RMK College of Engineering and Technology

Department
of
Electronics and Communication Engineering

CS 6551
Computer Networks
 Unit 1

Fundamentals
and
Link Layers
Prepared by

Jai Ganesh S
Asst.Professor - ECE
 Syllabus
• Building a network
• Requirements
• Layering and protocols
• Internet architecture
• Network software performance
• Link layer services
• Framing
• Flow control and Error Control
• Error detection
 Recalling the past….

• Before jumping in the syllabus lets revise

some fundamentals of computer
networks
• Characteristics of computer communication
systems.

• Components of data communication

• Transmission modes
1. Characteristics of computer
communication

Correct Timely
Delivery Delivery

Jitter
Accuracy
 2. Components of data communication
systems

MESSAGE

SENDER RECEIVER

TRANSMISSION MEDIUM
2. Components of data communication
systems
 3. Transmission Modes
Transmission
Modes

Uni
Bi directional
directional

Simplex Half Duplex

Full Duplex
 Getting Inside the syllabus

Fundamentals Syllabus
 Building a Network
• In order to understand how to build, operate and
program a network, a programmer has to explore
the below mentioned stages.

1. Explore the requirements of different

application and different communities.

2. Under the idea of network architecture.

3. Implementing the computer networks.

4. Evaluate the performance of computer

networks.
Requirements of a
network
 Requirements of a Network

• It is based on various aspects mentioned

below
1. Perspectives
2. Scalable connectivity
3. Cost effective resource sharing
4. Support for common services
5. Reliability
6. Manageability
 1.Based on perspective

Application Programmer Network operator Network Designer

They would list the They would list the They would list
services that are characteristics of the properties of
required by the the system that is cost effective
applications. easy to administer design.
and manage.
Ex: Each message sent by the Ex: network resources
application will be delivered Ex: Adding and are fairly utilized and
without error removing a machine fairly allocated to all
should be easy the users
 2.Scalable connectivity
Computers are connected in 2 major categories.
Directly connected Indirectly connected

• They are also called as point • They are also called as multiple

to point connection access.

• In this case more than two nodes

• The nodes are directly
may share a single physical link.
connected form end to end as
shown in the figure. • They are further classified in to
two types. They are,

1. Circuit switched networks

2. Packet switched networks

Circuit switched
• They are widely used in the
networks
telephone networks.

• It first establishes a dedicated

circuit across a sequence of links

• Then it allows source node to

send a stream of bits across this
circuit to a destination
 • These networks are widely used in most of the
Packet switched applications
networks
• Messages are transmitted as packets

• They follow store and forward concept of

transferring packets

• Packets may take any route and they are routed

by the intermediate routers.

• These packets are rearranged at the receiver side.

• Many discrete networks are interconnected

together to form a internetwork or INTERNET.

• Nodes are addressed logically so that they can be

identified by the neighbouring devices in the
network

• Packet transmission occurs in 3 modes, they are:

1. Unicast

2. Multicast

3. broadcast
3.Cost Effective Resource Sharing
• A resource in a network can be shared by
many users and it must be cost effective.

• This can be achieved by a concept called as

MULTIPLEXING.

• The figure shows the concept of multiplexing.

Data sent by corresponding source are
multiplexed on to single channel and they
are de multiplexed at the receiver side and
delivered to corresponding nodes.

• There are 2 types of multiplexing, they are:

1. Frequency division multiplexing [FDM]

2. Time division multiplexing [TDM]

 Types of Multiplexing
Frequency Div Mux Time Div Mux

• The channel is portioned in to time slots

• Total bandwidth of the channel is and each is allotted for particular user
divided into non overlapping for transmission
frequency sub bands and each user is
• They are further classified in to 2 types
allotted different sub band

1. Synchronous TDM

2. Asynchronous TDM
Synchronous TDM

• Each frame contains at

least one time slot for each
device.

• If the device do not have

any message to transmit ,
the slot remains free.

Asynchronous TDM

• The slot is allotted

dynamically according to
the activity of the devices.

• No possibilities for free

slot in the frame
 Resource sharing Contd…
• The size of the network also plays an important role in cost
effective resource sharing.

• Based on the size, the networks are classified as

PAN LAN MAN WAN

 Meant for only  Meant for  Meant for  Meant for long
1 person single office or town or city distance over
 Distance – 1 building  Distance - few a geographical
meter  Distance – few 10’s of K mts area
K mts  Distance – few
100’s of K mts
 4.Support for common services

• Different applications need different

requirements.

• An application designer may identify the

common services that applies for all the
applications and can build a common
channel that can be used by all other
applications.

• Any applications can be developed by using

the common channel
 5.Reliability
• Reliable message delivery is one important aspect
of networks.
• Its very difficult to provide reliability
without knowing the types of failures
that occurs in the network
3 cases of failure :

Packet lost Physical link is

Bit Errors Remedies : Down
 Error control Mechanism – they control the errors in the packets for case 1
and request for retransmission for case 2

 Congestion control mechanism – they are implemented to avoid the loss of

packets due to congestion in the network
 6.Manageability
• Making changes on the growing
network should not be a complex
task.

• Adding or removing any

new host should not affect
the other machines on the
network.

• Troubleshooting a network should

not get complicated in the process of
expanding the network.
 Final Verdict
• Thus a computer network must provide a
general, cost effective, fair and robust
connectivity among the group of
computers.

• It must also deal with the varying

environments to adopt the changes and
latest technologies.

• They must be manageable by any humans

with varying skill sets.
 End
of
Requirements
Network Architecture
 Network Architecture
• Building a network with all the pre
discussed requirements is not an easy job.
• To get rid of this complexity the networks
are given some architecture. They are
referred to as Blue Prints.
• There are 2 most widely used network
architectures. They are as follows
1. OSI Architecture
2. Internet Architecture
 Features of Architectures
• The architectural design provides 2 main
features, they are:
It decomposes the problem of building a
network into more components
- Several layers are added and each is assigned with
different functionality.

It provides a more modular design

- Making modifications becomes easier.
- If any new services wanted to be added , then the
modifications can be done only at 1 respective layer.
Other layers can be reused as such
Does Layering Really meets
the expectation..??
OSI Layer Model
 Introduction to OSI model
• OSI model is introduced by ISO (International Standard
Organization)

• It is a set of protocols that allows any two different systems

to communicate regardless their underlying architecture

• OSI contains 7 layers.

• They are grouped into 3 sub groups

Network Support layers
 Deals with physical aspects of moving data
from one device to another device

Transport Layer
 Ensures the reliable data transmission

User Support layers

 Deals with the interoperability
among unrelated software systems
Layers and their functions
 1. Physical Layer
• It co ordinates the functions
required to transmit a bit stream
over a physical channel. It also
deals with the mechanical and
electrical properties of the
interference and transmission
medium

Transmission Rate Physical Topology Bit

Synchronization
 2.Data Link Layer
• This layer is responsible for
sending group of packets to the
adjacent nodes without any
errors.

• This is called as node to node

delivery

 Framing
 Physical Addressing
 Flow Control
 Error Control
 Access Control
 3. Network Layer
• This layer is responsible for routing
the packets from one network to
another network.

• This is referred to as source to

destination delivery

Logical Addressing Routing

 4. Transport Layer
• This layer is responsible
for end to end delivery of
entire message (Not
packets or Frames)

• Functions:
• Service port addressing
• Segmentation and reassembly
• Connection control
• Connection oriented services
• Connection less services

• Flow control and error control

Segmentation and reassembly
Service Port addressing
 5. Session layer
• The main function
of this layer is to
establish, maintain,
and synchronize the
interaction between
two connected
nodes.
• Functions
• Dialogue control
• This function decides which application send / receive
data and connection type (Simplex or Duplex)

• Synchronization
• Maintains the same data rates at both sender and
receiver nodes.
 6. Presentation layer
• This layer is responsible
for presenting the data
in reliable format, as
different applications
uses different coding
standards
• Functions:
• Translation
• Encryption and
decryption
• Compression
Presentation layer
 7. Application Layer
• This layer provides the
end user interface for
network applications such
as mails, fund transfers,
Email Services
online shopping sites, etc.,

Browsers
Online
Shopping
Services
 End
of
OSI Model
Internet Architecture
 Internet Architecture

• They are also called as TCP / IP

Architecture or TCP / IP Protocol
Suite.

• This model uses various protocols at

different layers. They are discussed in
up coming slides.
TCP / IP Protocol suite
 1. Network interface layer

• The physical and datalink layer of the OSI

model are jointly known as Network
Interface layer in TCP / IP Model

• There is no specific protocols in this layers

and they supports all standards

• They particularly deals with the physical

connection between the nodes
 2. Network Layer

• They are also called as Internet layer.

• This layer defines the official packet

format.

• Protocol used is Internet Protocol (IP)

• They are responsible for successful

delivery of packets from one host to
another host
 Internet Protocol (IP)
• It is an transmission mechanism used for sending the
packets. It uses connection less protocol. IP in turn uses 4
protocols.
• Address Resolution Protocol (ARP):
 It is used to find the physical address of the node when
the logical address is known.
• Reverse Address resolution Protocol (RARP):
 It is used to find the logical address of the node when the
physical address is known.
• Internet Control Message Protocol (ICMP):
 It is an mechanism used by the hosts and gateways to
send notifications about the datagram problem to the
sender.
 It contains control and error messages.
• Internet Group Message Protocol (IGMP):
 It is used to facilitate the transmission of message to a
group of recipients (Multicast)
Address Resolution Protocol - ARP
Reverse Address Resolution Protocol -
RARP
ARP Vs RARP
Internet Control Message Protocol -
ICMP
Internet Group Message Protocol -
IGMP
 3. Transport layer
• They are represented by 3 protocols,

• Transmission Control Protocol (TCP):

 This is an reliable Connection Oriented protocol which allows the segments of
one machine to another machine without any error.
 A connection is established between the sender and receiver before the
transmission starts. This is called as handshaking.
 Each segments are numbered and on successful receipt of segments they are
acknowledged by the receiver.

• User Datagram Protocol (UDP):

 It is an unreliable , connection less protocol.
 No hand shaking process is done before transmission.
 No acknowledgement is received after delivery of segments.

• Stream Control Transmission Protocol (SCTP):

 They are used for newer applications such as Voice Over Internet
 This is simply the combination of both TCP and UDP
 4. Application Layer
• It is a combination of session layer, presentation
layer and application layers of OSI model.

• It supports various protocols.

TELNET

FTP – File Transfer Protocol

SMTP – Simple Mail Transfer Protocol

DNS – Domain Name System

HTTP – Hyper Text Transfer Protocol.

Telnet

This protocol allows

the user of one
machine to log on to
another machine
and can work from
here
FTP – File Transfer Protocol

This protocol
provides a
procedure to
move data from
one machine to
another machine
more efficiently
SMTP – Simple Mail Transfer
Protocol

This protocol is
mainly used for
mail transfers from
various servers.
DNS – Domain Name System

This protocol is
helpful in
mapping their
host name with
their network
address
HTTP – Hyper Text Transfer Protocol

This protocol is
used in fetching
the webpages on
the world wide
web and other
functionalities as
well
 End of
TCP / IP Model
&
End of
Network Architecture
 Network
Performance
 Network Performance
• One important aspect of a network is to
measure How good it is?..
• This is referred to as network
performance.
• There are 4 major performance
measuring parameters.
Bandwidth
Throughput
Latency
Bandwidth and delay product
 1. Bandwidth
• This is one important measure of a network.

• It can have 2 forms, they are:

Bandwidth in HERTZ:
It is the range of frequencies
contained in a signal which a
channel can pass

Bandwidth in BITS per Second:

It is the measure of number of
bits transferred over a network
in a certain period of time.
 2. Throughput
• Throughput is the rate of successful message
delivery over a communication channel.
• This decides how fast we can send the data.
• A link may have bandwidth of “B” bits per
second but we can transmit only “T” bits
per second trough this link.
• Where T is always less than B.
• A link may be a 1Mbps Line but we can
send data only at the speed of 200Kbps.
 3. Latency
• It is defines as “how long it takes for an entire message to
completely arrive at the destination”
• They are made out of 4 components:
 Propagation time
 Specific amount of time a signal can take to propagate from one end to
another end.
 Propagation time = distance / propagation speed
 Transmission time
 It is the amount of time that channel takes to transmit the unit of data.
 It measures the time from “First bit leaving the sender” to “Last bit
arriving the Receiver”
 Transmission time = size of the message / bandwidth
 Queuing time
 This is the time required for each intermediate node to hold the message
before it is processed.
 Processing time
 This is the time that the router take to process the header of the packet
 4. Bandwidth and delay product

• The measure of bandwidth and delay

gives the over all capacity of the link.

• This helps in computing the maximum

number of bits that a link can handle
at any instance.
 End
of
Network
Performance
Link Layer Services
 Link Layer Services

• The main functions of data link layer

are as follows,

1. Framing

2. Flow control

3. Error control

4. Error detection
 1. Framing
• It is the process in which the packets
received from the network layer are
divided into smaller parts called as
frames.

• There are 2 types of framing

Fixed length framing.
Variable length framing.

• The challenging task is to identify the

start and end of the frame.
 Fixed length framing

• In this method the size of the frame is

fixed and thus there is no need for
defining the boundaries of the frame.
 Variable Length Framing

• The size of the frame is not fixed.

(Varies).

• Hence we need a way to define the

frame boundaries.

• There are 3 approaches

Byte oriented framing

Bit oriented framing

Clock based framing

 Byte Oriented Framing
• Normally the frame from the upper layer comes with
the header and trailer information .

• Header contains source and destination address

• Trailer contains error detection and correction

information

• In order to identify the frame boundaries an 8 bit flag

is added at both the ends

• These flag may contain protocol dependent special

characters.
 Problems with Flags

• This method is generally suited for text transmission, but

in case of audio and video transmission any flag value can
be a part of the data.

• If the pattern is as same as the flag, then the receiver may

misinterpret that it is the end of the frame

• This problem can be addressed by BYTE Stuffing.

• In this method an extra byte is added to the data section if

the pattern is as same as the flag. These bytes are usually
referred to as Escape Characters (ESC)
Byte Stuffing
Example for Byte Stuffing
• The universal codes that we use now a
days may conflict with the 8 bit
characters of the flag.

• The Protocols that uses this approach are

BISYNC, PPP, DDCMP Etc.,

• Hence we move towards an another

approach called as Bit stuffing.
 Bit Oriented Framing
• In this type of framing flag uses an constant 8 bit
sequence “01111110” as the delimiter to indicate the
boundaries of the frames.

• These flags may also create the same problem as Byte

oriented protocols.

• To overcome this problem we stuff 1 bit to the data to

prevent the pattern from looking as same as flag.

• This strategy is called as BIT Stuffing.

• In this method, if 5 consecutive 1 bits
are encountered an extra 0 bit is
added regardless to the value of next
bit.

• These bits are eventfully removed at

the receiver side.

• The protocols that uses this approach

are SDLC, HDLC Etc.,
 Clock Based Framing
• The best example for clock based framing is SONET (Synchronous
Optical NETworks)

• They are widely used for optical communications.

• It addresses both framing and encoding problems.

• It also provides multiplexing several low speed telephone lines with

single High Speed telephone Line.

• The frame structure is shown in the figure.

• It contains 9 rows of 90 bytes each

• First 3 bytes of each row is overhead and the rest is

available for data

• The first 2 bytes of each frame contains the special bit

pattern and these pattern are used for determining the
start and stop of the frame by the receiver.

• The receiver checks for this bit pattern once in every

810 bytes (9*90 = 810 bytes)

• The overhead bytes are encoded using NRZ and the bit
pattern is scrambled.

• Scrambling is done by ExOR ing the data bits with well

known bit patterns.
 End
of
Framing
2. Flow Control
Traffic Example
 Flow Control
• It co ordinates the amount of data that can
be sent before receiving and
acknowledgement.
• They are widely used in the synchronizing
purpose. At any case the flow of transmitter
should not overwhelm the flow of receiver.
• The two main protocols that are used for flow
control are
Stop and wait Protocol
Sliding window Protocol
 Stop and Wait Protocol
• This is used for sending one frame at a time.

• They are simple and easy to implement

• This protocol sends one frame at a time and

waits for the acknowledgement (ACK) from
the receiver.

• Next frame is sent only if receiver

acknowledges the previous frame. If ACK is
not received then the transmission is stopped
temporarily.
Flow diagram for stop and wait protocol
 Merits and Demerits

• Merits
• Each frame is acknowledged individually

• No loss of frames during transmission

• Demerits
• Very inefficient

• Transmission is very slow.

 Sliding window Protocol
• This protocol is well suited for sending multiple frames.

• The receiver receives only some of the frames to

indicate the receipt of multiple frames.

• Each frame are numbered from 0 to N-1 as N-1 is the

size of the frame.
Sliding window form Sender and Receiver
Prespective
Example
 End
of
Flow Control
3. Error Control
 Error Control
• Most of the channels used in the real time
communications are noisy channels

• These noise in the channel causes errors in the

passing data.

• Thus a special mechanism is required to control

these errors.

• Error control Protocols are categorised as

Stop and wait ARQ

Sliding window ARQ

 Go back n ARQ
 Selective Repeat ARQ (Selective Reject ARQ)
 Stop and Wait ARQ
• ARQ – Automatic Repeat Request

• This protocol adds simple errorcontrol mechanism to

ordinary stop and wait protocol.

• After transmitting one frame the sender waits for an

acknowledgement before sending the next packet.

• If the ACK is not received within a predefined time period

then the frame is retransmitted by the sender.

• Various scenarios,

 Frame delivered , ACK received on time.

 Frame delivered, ACK is lost

 Frame is lost

 Frame Delivered , ACK Received after Time out.

Flow Diagram for different scenarios
 Drawback
• The major draw back in the stop and wait ARQ
is that, in case (b) and (d) the frame is
retransmitted even though the first frame is
delivered successfully.

• This creates duplicate copies of frames at the

receiving side.

• This can be overcome by adding sequence

numbers to the frames and acknowledgements.

• By this the receiver could identify either the

frame is duplicated frame or Original frame.
 Sliding window ARQ

• These protocols are designed to

transmit multiple frames at a time.

• They are further classified into 2 types,

Go Back n ARQ

Selective Repeat ARQ

 Go Back n ARQ
• In order to improve the channel efficiency
multiple frames are transmitted and a copy
is retained until they are acknowledged.

• The different scenarios are,

Damaged frame

Lost frame

Lost ACK
Lost Frame
Damaged Frame
Lost Ack
 Drawback of Go Back n ARQ

• In case of lost frame or lost ack all the

frames are retransmitted, which again
creates the duplication of frames at the
receiver side.

• Hence to overcome this we go for another

type called as Selective Repeat ARQ.
 Selective Repeat ARQ
• In this protocol only the specific damaged or
lost frames are retransmitted.
• This protocol can re arrange the sequence of
frames at the receiver based on the sequence
numbers.
• The flow diagram for various cases are shown
in the upcoming slides.
Damaged frame
Lost frame
Lost ACK
Flow Diagram – Selective Repeat
ARQ
 End
of
Error Control
4. Error Detection
 Error Detection
• For controlling the errors the errors are to be
identified first.

• Errors are introduced in to the channel due

to thermal noise and electrical interference.

• We hereby discuss 3 approaches for error

detection.

Two Dimensional Parity

Checksum

Cyclic Redundancy Check (CRC)

 1. Two Dimensional Parity

• It is simple and based on 1 Dimensional

parity, which usually involves adding
one extra bit to a 7 bit code to balance
the number of 1’s in the byte
• 2 D Parity also does the similar calculation for each bit
positions across each byte contained in the frame.

• This results in a extra byte as shown below.

• Consider Even parity

This can detect all 1,2,3 bit

errors and most of the 4
bit errors.

14 bits of redundant

information is added for a
42 bit message.
 2. Checksum
• These codes are based on addition

• This method uses summing algorithm.

• The idea behind this method is add up all the

words that are transmitted and then transmit the
result of that sum.

• This result is referred to as checksum.

• The receiver performs the same operation and

compares the result with the checksum
transmitted.

• Based on the result the receiver identifies the

presence of error.
 Sender Site Algorithm

The message is divided in to 16 bit words.

The value of checksum word is set to “0”

All words including checksum are added

using 1’s Complement addition

The sum is complemented to get checksum

The check sum is sent along with the data.

 Receiver site Algorithm

• The message is divided into 16 bit

words (including checksum)

• All words are added using 1’s

Complement addition

• The sum is complemented to get the

new check sum

• If the new checksum is “0” then the

message is accepted. Else it is rejected.
Example
Problem
• This alg is better than the parity bits
as the redundancy bits are 16 for any
length of the data

• They do not provide error detection if

one bit increase by some value and
other bit reduces by the same value.

• This alg is advantageous by their

simple design and easy to implement.
 3. Cyclic Redundancy Check (CRC)
• It is based on binary division

• The msg bits are represented by polynomials

using the values of each bit in the message as the
coefficient.
Example Problem
Versions of CRC
 End
of
Error Detection
 End
of
UNIT 1