Unit 4
Syllabus
Multiplexing
Switching: Circuit switched networks, Data gram networks, Virtual circuit networks
Dial up modems,
DSL
Error Detection and Correction: Block coding, cyclic codes, Linear block codes
Multiplexing
Needed when bandwidth requirements are more than available bandwidth
Multiplexing is set of techniques that allowsimultaneous
transmission ofmultiple signals acrosssingle data link
Multiplexing
Input lines are fed toMultiplexer(MUX).
MUX combines multiple input line to single stream of data
At output end, The stream is fed to aDemultiplexer(DEMUX)
Categories of Multiplexing
FDM
Analog multiplexing technique
For digital signals, they must be converted to analog signals first
FDM
Individual signals modulates different carrier frequencies
Each carrier frequency is separated by a guard band (unused bandwidth) to prevent signals from overlapping
In addition, choice of carrier frequencies must ensure that they do not interfere with data being carried
FDM: Modulation process
FDM: Demodulation process
Guard Bands
Application of FDM
AM broadcasting
AM radio = 530−1700KHz
Each radio station needs 10KHzofbandwidth
Each station uses different carrier frequency i.e. shifting its signal and multiplexes
This combined signals and transmitted Receiver filters the signals as per need
FM broadcasting
88−108MHz
To Do!
Simple numerical problems at Page 158−159 will give a better
WDM
WDM and FDM is similar since frequency and wavelength are related by the relation
λ= c
µ
Mostly used for optical fibers
Splitting and combining light sources are done using prisms
Time Division Multiplexing
Digital Multiplexing technique
Analog data should be digitized prior of TDM
Instead of sharing portion of bandwidth, time is shared
Synchronous TDM
In synchronous TDM scheme, each input connection has an allotment in the output,even if it is not sending data
Data flow much be chopped into aunit
Unit can be a bit, a character or block of data
Synchronous TDM example
If the input time slot is T seconds then output time slot is Tn seconds
wheren is number of channels
A round of data from all units is collected as a frame
Data rate of output link must be n times the data rate of input
connections to guarantee the flow of data
TDM numericals
Switching
A switched network consists of a series of interlinked nodes called
switches
Switches are devices capable of creating a temporary connection
Switching
Methods of switching
Switching and TCP/IP layers
Physical layer
Circuit switching only
Data-link layer
Packet switching Packet isframes/cells Virtual switching method
Network Layer
Packet Switching
Currently internet uses Datagram methods but its is being moved to Virtual switching method
Application layer
Circuit switching
Circuit Switching
Working of circuit switched network
Case: A→ M communication
Setup phase
A send request of connection to M
Must be accepted by all switches in path and M itself Hence acircuit/path is reserved on each link
Combination of dedicated circuits defines the dedicated transmission path
Data transfer phase
Data is transmitted along the defined circuits
Since transfer is at physical layer, so data is not packeted
No addressing involved during data transfer even though there is an end-to-end addressing for nodes during setup phase
Switches route data based on their occupied band (FDM) or time slot(TDM)
Tear-down phase
Efficiency
Low efficiency
Resourced are locked for the period of transfer
Time delay
Minimal Delay
Delay
Total delay = time for (creating connection + transfer data + disconnect the circuit)
There is no hold-up time at nodes
Packet Switching
In this technique, data is framed into packetsof fixed orvariable
sizes.
Size of packet is governed by network and governing protocol
No resource allocation for the packet
No reserved bandwidth
No time scheduling for the process
Resources are allocated on demand onfirst-come-first-served basis
When a switch receives a packet, if other apckets are being processed, it must wait.
Two types:
Datagram Networks
Each packet (=datagram) is treated individual of each other
Datagram Network
Works atnetwork layer
Switches in datagram network are traditionally calledrouters
As one can see from figure, packets may arrive at destination in
out-of-orderfashion.
Packets may also be lost of dropped because of a lack of resource
Connectionless network
Router does not keep information about their connection state There is no set-up and tear-down phase
Routing table
Each router has a routing table dynamic and updated periodically
Destination address
Every datagram contain header
header contains destination address
routers examines it and consults routing table to find appropriate port for forwarding
Destination address remains same all through the journey
Efficiency and Delay
Efficiency
Efficiency is better than circuit switched network as resources are allocated only on-demand
Delay
Although there are no set-up and tear-down phases, packets may experience long delay periods for availability of resources
Delay in datagram network
Virtual Switched Networks
Cross between circuit switched network and datagram network
Set-up, transfer and tear-down phase exist
Resources can be allocated either during setup phase (similar to circuit-switched) or on-demand (datagram).
Data is packeted carrying address in header Two types of addressing
Local addressing= local jurisdiction
Global addressing= end-to-end jurisdiction
Address has local jurisdiction only i.e. it just has information about next switch and port
All packets follow the same path established during connection (same as circuit switched network)
Virtual Switched Networks
Virtual Switched Identifier
Unlike global address, VCI is just a small number used by a frame between two switches
At arrival and during departure, VCI’s are different
Three phases
Set-up phase
Data transfer phase
Source to destination data transfer
Source to destination data transfer
Suppose end-to-end transmission isA→B
Two steps:
Setup request Acknowledgment
Set-up request
Source A sends a setup frame to switch 1 Switch 1 receives frame and analyzes
Frame is goingA→B through port 3
Switch creates an entry in its table for virtual circuit and fills three of four data points:
Incoming port (1)
Chooses available incoming VCI outgoing port (3)
Frame is forwarded to switch 2 through port 3 and it repeats the steps
Acknowledgement
Process of setup phase
Process of setup phase
Process of data transfer
Tear-down phase
After sending all frames, Source A sends a special frame called a
tear-down request
Destination B responds with a tear-down confirmation frame
Delay and efficiency
MODEMS
MODEM stands for Modulator-Demodulator
A modulator creates a bandpass analog signal from binary data
Dial-up MODEMS
Telephone lines can carry voice data for frequencies between
300−3300Hz, thus a bandwidth of 3000Hz
600−3000Hz i.e a bandwidth of 2400Hz can be used for data
MODEMS are used to transmit digital data over telephone lines
MODEM usage in communication
Input computer sendsdigital data to modulator in MODEM which
converts it inanalogform
Analog data is communicated using telephone line.
Output MODEM demodulates this analog data todigital data nd
feeds it to output computer
Communication can be bi-directional
DSL
When MODEM based communication reached its peak data rate, telephone companies devised new technology termed DSL (Digital Subscriber Line)
DSL is a set of technologies like ADSL, VDSL, HDSL and SDSL.
ADSL = Asymmetric DSL
The term asymmetric essentially indicates that data rate in
downstream is more than data rate for upsteam direction of data flow. It uses existing local loop (existing telephone lines)
ADSL
Twisted pair cables used in telephone lines are capable of 1.104MHz
Filters are usually installed limiting the bandwidth to 4KHz which is
sufficient for voice communication
These filters are removed in this case
1.104 is subdivided into 3 channels as follows:
0−4Kz = voice communication 26−108KHz Upstream
138−1104KHz Downstream
ADSL channels
ADSL
ADSL
Error Detection and correction
Errors are changes in bit patterns at generation, during transmission and/or during interpretation
Two types:
Single-bit error
Only one bit of a given data (byte, character or packet) is corrupted where 0 and 1 gets interchanged
Burst error
2 or more bits change their bit values
Redundancy
Some extra bits are added by sender and removed by receiver Redundant bits are added within data bits according to a rule This rule is shared between sender and receiver
Types of errors
Error Coding
Redundancy is achieved via various coding schemes Coding schemes are broadly divided in two categories:
Block coding Convolution coding
Block Coding
Message is divided into blocks (of say k bits)
k bit long blocks are called datawords
r redundant bits are added to dataword to make its lengthn=k+r
Block Coding
Basic idea:
k bit dataword makes 2k dataword combinations
n bit codeword makes 2n codeword combinations
n>k ⇒number of possible codewords are more than number of
possible datawords
Block coding process isone-to-onei.esame dataword is always
encoded as same codeword
2n−2k codewords are never used
These codewords are called invalidsor illegal
Error detection includes:
Block coding
Block coding
Hamming distance
Hamming distance between two wordsof same size is the number of
differences between corresponding bits
It is written as d(x,y)
It can be found by applying XOR operator ⊕
Examples:
d(111,100) = 111⊕100 = 011 so hamming distance is 2 (number of
1’s)
Hamming distance and error detection
Hamming distance gives an idea about number of corrupted bits within transmission
In a set of codewords, dmin represents the minimum hamming
distance between all possible pair of codes
Lets say we can detect upto s errors
Ifs errors occurred during transmission,dmin=s
If our system detects upto s errors, minimum distance between valid
codewords must be s+ 1 so that they do not match with valid
codewords
Rephrasing, if the minimum distance between all valid codewords is
s+ 1, the received codeword cannot be erroneously mistaken for
Linear block code
A linear block code is one in which the XOR (addition modulo-2) of two valid codeword creates another codeword
Example
S. No. Dataword Codeword
1 00 000
2 01 011
3 10 101
4 11 110
Table: Linear coding scheme where similar Dataword bits adds 0 and dis-similar Data word bits adds 1
Linear block code
dmin is number of 1’s in the nonzero valid codeword with smallest
number of 1’s
Parity Check
It is a linear block code
If dataword isk bit thenn =k+ 1
Parity bit is added tomake the total number of 1’s even
Parity Check Example
S. No. Dataword Codeword
1 0000 00000
2 0001 00011
3 0010 00101
4 0011 00110
... ... ...
16 1111 11110
Table:Parity Check example
Graphical representation
Generation of parity bit
Generator takes a copy of 4 bit dataword ((a0,a1,a2,a3))
Generates a parity bitr by adding (modulo-2) them
r =a3+a2+a1+a0
r is zero if number of 1’s are even and 0 otherwise
At reception, same procedure is performed on 5 bits
The result is calledSyndrome
s0 =b3+b2+b1+b0+q0
s0 = 0 when number of 1’s in received codeword is even and 1
otherwise
s0 is passed to decision logic analyzer. If s0 ⇒No detectable error
Cyclic codes
Cyclic codes are linear codes with an extra feature: if a codeword is cyclically shifted then it results in another codeword
Ex: If 1011000 is cyclically shifted then it results in 0110001 which is also a codeword
If first word is a0 to a6 and second word isb0 to b6 then shifting means:
b1=a0
b2=a1
b3=a2
b4=a3 b5=a4
b6=a5
Division in CRC encoder
CRC
Checksum
Its an error detecting technique which can be applied to a message of any length
Its used by network and transport layers rather than data-link layer Method:
Message of divided intom-bit units
Generator produces extram-bit unit calledchecksum
Receiver creates a new checksum
If the new checksum is all 0’s, message is accepted and discarded otherwise
Practically, checksum need not be added at the end of message, it can be inserted in between too.
