Direct Link Networks - Encoding

Similar documents
Chapter 2: Sample Questions, Problems and Solutions Bölüm 2: Örnek Sorular, Problemler ve Çözümleri

ECE 333: Introduction to Communication Networks Fall Lecture 4: Physical layer II

Lecture 13: DUART serial I/O, part I

Objectives. Some Basic Terms. Analog and Digital Signals. Analog-to-digital conversion. Parameters of ADC process: Related terms

Analysis of SDR GNSS Using MATLAB

Encode Decode Sample Quantize [ ] [ ]

E X P E R I M E N T 13

Sampling. Introduction to Digital Data Acquisition: Physical world is analog CSE/EE Digital systems need to

General Model :Algorithms in the Real World. Applications. Block Codes

A New Space-Repetition Code Based on One Bit Feedback Compared to Alamouti Space-Time Code

Geometrical Optics Fiber optics

APPLICATION NOTE UNDERSTANDING EFFECTIVE BITS

Introduction to Wireless Communication Systems ECE 476/ECE 501C/CS 513 Winter 2003

Measurement of Equivalent Input Distortion AN 20

Lecture 4: Frequency Reuse Concepts

By: Pinank Shah. Date : 03/22/2006

Lecture 5 Transmission

4. INTERSYMBOL INTERFERENCE

Lecture 5 Transmission. Physical and Datalink Layers: 3 Lectures

Spread Spectrum Signal for Digital Communications

Intermediate Information Structures

Design of FPGA- Based SPWM Single Phase Full-Bridge Inverter

Electronic motor protection relay

Fitting Signals into Given Spectrum Modulation Methods

Comparison of Frequency Offset Estimation Methods for OFDM Burst Transmission in the Selective Fading Channels

Embedded Microcomputer Systems Lecture 9.1

PRACTICAL FILTER DESIGN & IMPLEMENTATION LAB

Lecture Progression. Followed by more detail on: Quality of service, Security (VPN, SSL) Computer Networks 2

Solution 2 Discussion:

CSE 461 Bits and Links. David Wetherall

SELEX Elsag. 5/18/2012 R. Pucci SDR 12 WinnComm 1

LAB 7: Refractive index, geodesic lenses and leaky wave antennas

TO DETERMINE THE NUMERICAL APERTURE OF A GIVEN OPTICAL FIBER. 2. Sunil Kumar 3. Varun Sharma 4. Jaswinder Singh

3. Error Correcting Codes

Lecture 2: Links and Signaling"

CHAPTER 5 A NEAR-LOSSLESS RUN-LENGTH CODER

Subscriber Pulse Metering (SPM) Detection

CSEP 561 Bits and Links. David Wetherall

INCREASE OF STRAIN GAGE OUTPUT VOLTAGE SIGNALS ACCURACY USING VIRTUAL INSTRUMENT WITH HARMONIC EXCITATION

Computer Networks

Beam Pattern Scanning (BPS) versus Space-Time Block Coding (STBC) and Space-Time Trellis Coding (STTC)

What is Multiple Access? Code Division Multiple Access for Wireless Communications. Time Division Multiple Access (TDMA)

x y z HD(x, y) + HD(y, z) HD(x, z)

Lecture 4. Antennas, db, and Introduction to Radio Propagation

Outline / Wireless Networks and Applications Lecture 3: Physical Layer Signals, Modulation, Multiplexing. Cartoon View 1 A Wave of Energy

CSE 561 Bits and Links. David Wetherall

Announcements : Wireless Networks Lecture 3: Physical Layer. Bird s Eye View. Outline. Page 1

COMPRESSION OF TRANSMULTIPLEXED ACOUSTIC SIGNALS

Lecture Progression. Followed by more detail on: Quality of service, Security (VPN, SSL) Computer Networks 2

Lecture 3: Data Transmission

(2) The MOSFET. Review of. Learning Outcome. (Metal-Oxide-Semiconductor Field Effect Transistor) 2.0) Field Effect Transistor (FET)

Modulation Analysis for Wireless Infrared Communications

Physical Layer. Networked Systems (H) Lecture 3

CSE 123: Computer Networks Alex C. Snoeren. Project 1 out Today, due 10/26!

Physical Layer. Networks: Physical Layer 1

CPSC Network Programming. How do computers really communicate?

Peer-to-Peer Protocols and Data Link Layer

A Novel Small Signal Power Line Quality Measurement System

Single Bit DACs in a Nutshell. Part I DAC Basics

Zonerich AB-T88. MINI Thermal Printer COMMAND SPECIFICATION. Zonerich Computer Equipments Co.,Ltd MANUAL REVISION EN 1.

MEASUREMENT AND CONTORL OF TOTAL HARMONIC DISTORTION IN FREQUENCY RANGE 0,02-10KHZ.

A New Design of Log-Periodic Dipole Array (LPDA) Antenna

Introduction to CPM-OFDM: An Energy Efficient Multiple Access Transmission Scheme

INF 5460 Electronic noise Estimates and countermeasures. Lecture 11 (Mot 8) Sensors Practical examples

ECC Recommendation (14)02

HOW BAD RECEIVER COORDINATES CAN AFFECT GPS TIMING

Outline. Motivation. Analog Functional Testing in Mixed-Signal Systems. Motivation and Background. Built-In Self-Test Architecture

PERFORMANCE OF MULTIBEAM MIMO FOR NLOS MILLIMETER WAVE INDOOR COMMUNICATION SYSTEMS CHAPTER 6

Chapter 2. Physical Layer

Subband Coding of Speech Signals Using Decimation and Interpolation

Last Time. Transferring Information. Today (& Tomorrow (& Tmrw)) Application Layer Example Protocols ftp http Performance.

Chapter 3 Digital Logic Structures

Lossless image compression Using Hashing (using collision resolution) Amritpal Singh 1 and Rachna rajpoot 2

Physical Layer. Transfers bits through signals overs links Wires etc. carry analog signals We want to send digital bits. Signal

Chapter 3. GPS Signal Propagation Characteristics and its Modeling

The Potential of Dynamic Power and Sub-carrier Assignments in Multi-User OFDM-FDMA Cells

Resolution. Learning Objectives. What are the four types of resolution?

Compression Programs. Compression Outline. Multimedia. Lossless vs. Lossy. Encoding/Decoding. Analysis of Algorithms

GENERATE AND MEASURE STANDING SOUND WAVES IN KUNDT S TUBE.

Design of FPGA Based SPWM Single Phase Inverter

The Throughput and Delay Trade-off of Wireless Ad-hoc Networks

Massachusetts Institute of Technology Dept. of Electrical Engineering and Computer Science Fall Semester, Introduction to EECS 2.

The Potential of Dynamic Power and Sub-carrier Assignments in Multi-User OFDM-FDMA Cells

Methods to Reduce Arc-Flash Hazards

The Physical Layer Outline

PERFORMANCE COMPARISON BETWEEN MC-CDMA AND DS-CDMA SYSTEMS FOR AWGN AND RAYLEIGH FADING CHANNEL

}, how many different strings of length n 1 exist? }, how many different strings of length n 2 exist that contain at least one a 1

A study on the efficient compression algorithm of the voice/data integrated multiplexer

Announcement : Wireless Networks Lecture 3: Physical Layer. A Reminder about Prerequisites. Outline. Page 1

信號與系統 Signals and Systems

A GHz Constant KVCO Low Phase Noise LC-VCO and an Optimized Automatic Frequency Calibrator Applied in PLL Frequency Synthesizer

Cooperative Diversity Based on Code Superposition

IEEE Protocol Implementation And Measurement Of Current Consumption by Rajan Rai

信號與系統 Signals and Systems

CDS 270-2: Lecture 6-3 Optimum Receiver Design for Estimation over Wireless Links

EECS 122: Introduction to Computer Networks Encoding and Framing. Questions

ASample of an XML stream is:

SIDELOBE SUPPRESSION IN OFDM SYSTEMS

BANDWIDTH AND GAIN ENHANCEMENT OF MULTIBAND FRACTAL ANTENNA BASED ON THE SIERPINSKI CARPET GEOMETRY

CCD Image Processing: Issues & Solutions

Transcription:

Direct Lik Networks - Ecodig Readig: Peterso ad Davie, Chapter 2 Sprig 2018 CS 438 Staff, Uiversity of Illiois 1

Where are we? Applicatio Layers BitTorret (P2P) HTTP (Web) Skype (VOIP) IPTV (streamig media) Trasport TCP UDP Network IP Today Data Lik Physical Etheret 4G DSL WiFi Sprig 2018 CS 438 Staff, Uiversity of Illiois 2

Direct Lik Networks All hosts are directly coected by a physical medium Key poits Ecodig ad Modulatio Framig Error Detectio Medium Access Cotrol Sprig 2018 CS 438 Staff, Uiversity of Illiois 3

Iteret Protocols Ecodig Framig, error detectio, medium access cotrol Applicatio Presetatio User-level software Sessio Trasport Network Kerel software (device driver) Data Lik Physical Hardware (etwork adapter) Sprig 2018 CS 438 Staff, Uiversity of Illiois 4

Direct Lik Networks - Outlie Hardware buildig blocks Ecodig Framig Error detectio Multiple access media (MAC examples) Network adapters Sprig 2018 CS 438 Staff, Uiversity of Illiois 5

Hardware Buildig Blocks Nodes Hosts: geeral purpose computers Switches: typically special purpose hardware Routers: varied Sprig 2018 CS 438 Staff, Uiversity of Illiois 6

Nodes: Workstatio Architecture Fiite memory Scarce resource Rus at memory speeds, NOT processor speeds Processor Cache memory bus (MBUS) Memory iput/output bus (I/O BUS) NETWORK Network Adaptor Sprig 2018 CS 438 Staff, Uiversity of Illiois 7

Hardware Buildig Blocks Liks Physical medium Copper wire with electroic sigalig Glass fiber with optical sigalig Wireless with electromagetic (radio, ifrared, microwave) sigalig Two cups ad a strig Sprig 2018 CS 438 Staff, Uiversity of Illiois 8

Liks - Copper Copper-based Media more twists, less crosstalk, better sigal over loger distaces Category 5/6 Twisted Pair 10-1Gbps 100m ThiNet Coaxial Cable 10-100Mbps 200m ThickNet Coaxial Cable 10-100Mbps 500m twisted pair coaxial cable (coax) copper core isulatio braided outer coductor outer isulatio More expesive tha twisted pair High badwidth ad excellet oise immuity Sprig 2018 CS 438 Staff, Uiversity of Illiois 9

Liks - Optical Optical Media Multimode Fiber 100Gbps 2km Sigle Mode Fiber 100-2400Mbps 40km optical fiber glass core (the fiber) glass claddig plastic jacket Sprig 2018 CS 438 Staff, Uiversity of Illiois 10

Liks - Optical Sigle mode fiber Expesive to drive (Lasers) Lower atteuatio (loger distaces) 0.5 db/km Lower dispersio (higher data rates) Multimode fiber Cheap to drive (LED s) Higher atteuatio Easier to termiate core of sigle mode fiber ~1 wavelegth thick = ~1 micro core of multimode fiber (same frequecy; colors for clarity) O(100 micros) thick Sprig 2018 CS 438 Staff, Uiversity of Illiois 11

Liks - Optical Advatages of optical commuicatio Higher badwidths Superior (lower) atteuatio properties Immue from electromagetic iterferece No crosstalk betwee fibers Thi, lightweight, ad cheap (the fiber, ot the optical-electrical iterfaces) Sprig 2018 CS 438 Staff, Uiversity of Illiois 12

Liks - Wireless Path loss Sigal atteuatio as a fuctio of distace Sigal-to-oise ratio (SNR Sigal Power/Noise Power) decreases, make sigal urecoverable Multipath propagatio Sigal reflects off surfaces, effectively causig self-iterferece Iteral iterferece (from other users) Hosts withi rage of each other collide with oe aother s trasmissio Exteral iterferece Microwave is tured o ad blocks your sigal Sprig 2018 CS 438 Staff, Uiversity of Illiois 13

Wireless Path Loss Sigal power atteuates by about ~r 2 factor for omi-directioal ateas i free space r is the distace betwee the seder ad the receiver The expoet i the factor is differet depedig o placemet of ateas Less tha 2 for directioal ateas Faster atteuatio Expoet > 2 whe ateas are placed o the groud Sigal bouces off the groud ad reduces the power of the sigal Sprig 2018 CS 438 Staff, Uiversity of Illiois 14

Wireless Multipath Effects Ceilig S Floor S Sigals bouce off surfaces ad iterfere with oe aother What if sigals are out of phase? Orthogoal sigals cacel each other ad othig is received! Sprig 2018 CS 438 Staff, Uiversity of Illiois 15

What is a Wireless Lik? Sprig 2018 CS 438 Staff, Uiversity of Illiois 16

What is a Wireless Lik? Sprig 2018 CS 438 Staff, Uiversity of Illiois 17

Wireless Bit Errors The lower the SNR (Sigal/Noise) the higher the Bit Error Rate (BER) How ca we deal with this? Make the sigal stroger Why is this ot always a good idea? Icreased sigal stregth requires more power Icreases the iterferece rage of the seder, so you iterfere with more odes aroud you Error correctio ca correct some problems Sprig 2018 CS 438 Staff, Uiversity of Illiois 18

Ecodig digital data (a strig of symbols) modulator a strig of sigals demodulator digital data (a strig of symbols) Problems with sigal trasmissio Atteuatio: Sigal power absorbed by medium Dispersio: A discrete sigal spreads i space Noise: Radom backgroud sigals Sprig 2018 CS 438 Staff, Uiversity of Illiois 19

How ca two hosts commuicate? 0.7 Volts -0.7 Volts Ecode iformatio o modulated Carrier sigal Phase, frequecy, ad/or amplitude modulatio Etheret: self-clockig Machester codig Techologies: copper, optical, wireless Sprig 2018 CS 438 Staff, Uiversity of Illiois 20

Ecodig Goal Uderstad how to coect odes i such a way that bits ca be trasmitted from oe ode to aother Idea The physical medium is used to propagate sigals Modulate electromagetic waves Vary voltage, frequecy, wavelegth Data is ecoded i the sigal Sprig 2018 CS 438 Staff, Uiversity of Illiois 21

Bauds ad Bits Baud rate Number of physical symbols trasmitted per secod Bit rate Actual umber of data bits trasmitted per secod Relatioship Depeds o the umber of bits ecoded i each symbol Sprig 2018 CS 438 Staff, Uiversity of Illiois 22

Aalog vs. Digital Trasmissio Aalog ad digital correspod roughly to cotiuous ad discrete Data: etities that covey meaig Aalog: cotiuously varyig patters of itesity (e.g., voice ad video) Digital: discrete values (e.g., itegers, ASCII text) Sigals: electric or electromagetic ecodig of data Aalog: cotiuously varyig electromagetic wave May be propagated over a variety of media Digital: sequece of voltage pulses May be trasmitted over a wire medium Sprig 2018 CS 438 Staff, Uiversity of Illiois 23

Aalog vs. Digital Trasmissio Advatages of digital trasmissio over aalog Cheaper Simpler for multiplexig distict data types (audio, video, e-mail, etc.) Easier to ecrypt Two examples based o modulator-demodulators (modems) Electroic Idustries Associatio (EIA) stadard: RS-232 Iteratioal Telecommuicatios Uio (ITU) V.32 9600 bps modem stadard Sprig 2018 CS 438 Staff, Uiversity of Illiois 24

RS-232 Commuicatio betwee computer ad modem Uses two voltage levels (+15V, -15V), a biary voltage ecodig Data rate limited to 19.2 kbps (RS-232-C); raised to 115,200 kbps i later stadards Characteristics Serial Oe sigalig wire, oe bit at a time Asychroous Lie ca be idle, clock geerated from data Character-based Sed data i 7- or 8-bit characters Sprig 2018 CS 438 Staff, Uiversity of Illiois 25

RS-232 Timig Diagram +15 Oe bit per clock tick Voltage + Voltage ever returs to 0V 0V is a dead/discoected lie -15 idle start 1 0 0 1 1 0 0 stop idle -15V is both idle ad 1 Time Sprig 2018 CS 438 Staff, Uiversity of Illiois 26

RS-232 Iitiate sed by Push to 15V for oe clock (start bit) Miimum delay betwee character trasmissios Idle for oe clock at -15V (stop bit) Oe character 0, 1 or 2 voltage trasitios Total Bits 9 bits for 7 bits of data (78% efficiet) Start ad stop bits also provide framig Sprig 2018 CS 438 Staff, Uiversity of Illiois 27

RS-232 Timig Diagram +15 Voltage + -15 idle start 1 0 0 1 1 0 0 stop idle Time Sprig 2018 CS 438 Staff, Uiversity of Illiois 28

Voltage Ecodig Biary voltage ecodig Doe with RS-232 example Geeralize before cotiuig with V.32 (ot a biary voltage ecodig) Commo biary voltage ecodigs No-retur to zero (NRZ) NRZ iverted (NRZI) Machester (used by IEEE 802.3 10 Mbps Etheret) 4B/5B Sprig 2018 CS 438 Staff, Uiversity of Illiois 29

No-Retur to Zero (NRZ) Sigal to Data High ð 1 Low ð 0 Commets Trasitios maitai clock sychroizatio Log strigs of 0s cofused with o sigal Log strigs of 1s causes baselie wader Both ihibit clock recovery Bits 0 0 1 0 1 1 1 1 0 1 0 0 0 0 1 0 NRZ Sprig 2018 CS 438 Staff, Uiversity of Illiois 30

No-Retur to Zero Iverted (NRZI) Sigal to Data Trasitio ð 1 Maitai ð 0 Commets Solves series of 1s, but ot 0s Bits 0 0 1 0 1 1 1 1 0 1 0 0 0 0 1 0 NRZ NRZI Sprig 2018 CS 438 Staff, Uiversity of Illiois 31

Machester Ecodig Sigal to Data XOR NRZ data with clock High to low trasitio ð 1 Low to high trasitio ð 0 Commets (used by IEEE 802.3 10 Mbps Etheret) Solves clock recovery problem Oly 50% efficiet ( ½ bit per trasitio) Bits 0 0 1 0 1 1 1 1 0 1 0 0 0 0 1 0 NRZ Clock Machester Sprig 2018 CS 438 Staff, Uiversity of Illiois 32

4B/5B Sigal to Data Ecode every 4 cosecutive bits as a 5 bit symbol Symbols At most 1 leadig 0 At most 2 trailig 0s Never more tha 3 cosecutive 0s Trasmit with NRZI Commets 16 of 32 possible codes used for data At least two trasitios for each code 80% efficiet Sprig 2018 CS 438 Staff, Uiversity of Illiois 33

4B/5B Data Symbols At most 1 leadig 0 0000 Þ 11110 0001 Þ 01001 0010 Þ 10100 0011 Þ 10101 0100 Þ 01010 0101 Þ 01011 0110 Þ 01110 0111 Þ 01111 At most 2 trailig 0s 1000 Þ 10010 1001 Þ 10011 1010 Þ 10110 1011 Þ 10111 1100 Þ 11010 1101 Þ 11011 1110 Þ 11100 1111 Þ 11101 Sprig 2018 CS 438 Staff, Uiversity of Illiois 34

4B/5B Cotrol Symbols 11111 Þ idle 11000 Þ start of stream 1 10001 Þ start of stream 2 01101 Þ ed of stream 1 00111 Þ ed of stream 2 00100 Þ trasmit error Other Þ ivalid Sprig 2018 CS 438 Staff, Uiversity of Illiois 35

Biary Voltage Ecodigs Problem with biary voltage (square wave) ecodigs Very wide (Ifiite) frequecy rage required, implyig Sigificat dispersio Ueve atteuatio Prefer to use a arrower frequecy bad 20 10 Voltage 0-10 -20-30 start 1 1 0 0 0 0 1 stop Sprig 2018 CS 438 Staff, Uiversity of Illiois 36

Biary Voltage Ecodigs Problem with biary voltage (square wave) ecodigs Very wide (Ifiite) frequecy rage required, implyig Sigificat dispersio Ueve atteuatio Prefer to use a arrower frequecy bad Types of modulatio Amplitude (AM) Frequecy (FM) Phase/phase shift Combiatios of these Sprig 2018 CS 438 Staff, Uiversity of Illiois 37

Example: AM/FM for cotiuous sigal Origial sigal Amplitude modulatio Frequecy modulatio Sprig 2018 CS 438 Staff, Uiversity of Illiois 38

Amplitude Modulatio idle 1 0 Sprig 2018 CS 438 Staff, Uiversity of Illiois 39

Frequecy Modulatio idle 1 0 Sprig 2018 CS 438 Staff, Uiversity of Illiois 40

Phase Modulatio idle 1 0 Sprig 2018 CS 438 Staff, Uiversity of Illiois 41

Phase Modulatio phase shift i carrier frequecy 108º differece i phase collapse for 108º shift Sprig 2018 CS 438 Staff, Uiversity of Illiois 42

Phase Modulatio Algorithm Sed carrier frequecy for oe period Perform phase shift Shift value ecodes symbol 135º 8-symbol example 90º 45º Value i rage [0, 360º) Multiple values for multiple symbols Represet as circle 180º 225º 270º 0º 315º Sprig 2018 CS 438 Staff, Uiversity of Illiois 43

V.32 9600 bps Commuicatio betwee modems Aalog phoe lie Uses a combiatio of amplitude ad phase modulatio Kow as Quadrature Amplitude Modulatio (QAM) Seds oe of 16 sigals each clock cycle Sprig 2018 CS 438 Staff, Uiversity of Illiois 44

Costellatio Patter for V.32 QAM 15º 45º Same algorithm as phase modulatio Ca also chage sigal amplitude 2-dimesioal represetatio Agle is phase shift 16-symbol example Radial distace is ew amplitude Sprig 2018 CS 438 Staff, Uiversity of Illiois 45

Example costellatio Sprig 2018 CS 438 Staff, Uiversity of Illiois 46

Commets o V.32 V.32 trasmits at 2400 baud i.e., 2,400 symbols per secod Each symbol cotais log 2 16 = 4 bits Data rate 4 x 2400 = 9600 bps Poits i costellatio diagram Chose to maximize error detectio Process called trellis codig Sprig 2018 CS 438 Staff, Uiversity of Illiois 47

Modulatio (Baud) Rate 5 bits = 5µsec A stream of biary 1s at 1 Mbps NRZI 1 1 1 1 1 What is a bit? Machester 1 bit = 1µsec 1 sigal elemet = 1µsec What is a sigal elemet? 1 bit = 1µsec 1 sigal elemet = 0.5µsec Sprig 2018 CS 438 Staff, Uiversity of Illiois 48

Modulatio (Baud) Rate 5 bits = 5µsec A stream of biary 1s at 1 Mbps NRZI What is the data rate? Data Rate (R) = bits/sec = 1 Mbps for both Machester 1 1 1 1 1 1 bit = 1µsec 1 bit = 1µsec 1 sigal elemet = 1µsec 1 sigal elemet = 0.5µsec What is the modulatio rate? Modulatio Rate = Baud Rate = Rate at which sigal elemets are geerated = R (NRZI) = 2R (Machester) Sprig 2018 CS 438 Staff, Uiversity of Illiois 49

Samplig Suppose you have the followig 1Hz sigal beig received How fast to sample, to capture the sigal? Sprig 2018 CS 438 Staff, Uiversity of Illiois 50

Samplig Samplig a 1 Hz sigal at 2 Hz is eough Captures every peak ad trough Sprig 2018 CS 438 Staff, Uiversity of Illiois 51

Samplig Samplig a 1 Hz sigal at 3 Hz is also eough I fact, more tha eough samples to capture variatio i sigal Sprig 2018 CS 438 Staff, Uiversity of Illiois 52

Samplig Samplig a 1 Hz sigal at 1.5 Hz is ot eough Why? Sprig 2018 CS 438 Staff, Uiversity of Illiois 53

Samplig Samplig a 1 Hz sigal at 1.5 Hz is ot eough Cat distiguish betwee multiple possible sigals Problem kow as aliasig Sprig 2018 CS 438 Staff, Uiversity of Illiois 54

What about more complex sigals? Fourier s theorem Ay cotiuous sigal ca be decomposed ito a sum of sies ad cosies at differet frequecies Example: Sum of 1 Hz, 2 Hz, ad 3 Hz sies How fast to sample? Sprig 2018 CS 438 Staff, Uiversity of Illiois 55

What about more complex sigals? Fourier s theorem Ay cotiuous sigal ca be decomposed ito a sum of sies ad cosies at differet frequecies Example: Sum of 1 Hz, 2 Hz, ad 3 Hz sies How fast to sample? --> aswer: 6 Hz Sprig 2018 CS 438 Staff, Uiversity of Illiois 56

Geeralizig the Examples What limits baud rate? What data rate ca a chael sustai? How is data rate related to badwidth? How does oise affect these bouds? What else ca limit maximum data rate? Sprig 2018 CS 438 Staff, Uiversity of Illiois 57

What Limits Baud Rate? Baud rate Typically limited by electrical sigalig properties Chagig voltages takes time No matter how small the voltage or how short the wire Electroics Slow compared to optics Baud rate Ca be as high as twice the badwidth of commuicatio Sprig 2018 CS 438 Staff, Uiversity of Illiois 58

What Data Rate ca a Chael Sustai? How is Data Rate Related to Badwidth? Trasmittig N distict sigals over a oiseless chael with badwidth B, we ca achieve at most a data rate of Number of sigals per secod 2B log 2 N Number of bits per sigal Nyquist s Samplig Theorem (H. Nyquist, 1920 s) Sprig 2018 CS 438 Staff, Uiversity of Illiois 59

What Data Rate ca a Chael Sustai? How is Data Rate Related to Badwidth? Trasmittig N distict sigals over a oiseless chael with badwidth B, we ca achieve at most a data rate of Number of sigals per secod Baud rate 2B log 2 N Number of physical symbols trasmitted per secod Bit rate Number of bits per sigal Nyquist s Samplig Theorem (H. Nyquist, 1920 s) Actual umber of data bits trasmitted per secod Relatioship Depeds o the umber of bits ecoded i each symbol Sprig 2018 CS 438 Staff, Uiversity of Illiois 60

Noiseless Capacity Nyquist s theorem: 2B log 2 N Example 1: samplig rate of a phoe lie B = 4000 Hz 2B = 8000 samples/sec. sample every 125 microsecods Sprig 2018 CS 438 Staff, Uiversity of Illiois 61

Noiseless Capacity Nyquist s theorem: 2B log 2 N Example 2: oiseless capacity B = 1200 Hz N = each pulse ecodes 16 symbols C = Sprig 2018 CS 438 Staff, Uiversity of Illiois 62

Noiseless Capacity Nyquist s theorem: 2B log 2 N Example 2: oiseless capacity B = 1200 Hz N = each pulse ecodes 16 symbols C = 2B log 2 (N) = D x log 2 (N) = 2400 x 4 = 9600 bps Sprig 2018 CS 438 Staff, Uiversity of Illiois 63

How does Noise affect these Bouds? Noise Blurs the symbols, reducig the umber of symbols that ca be reliably distiguished Claude Shao (1948) Exteded Nyquist s work to chaels with additive white Gaussia oise (a good model for thermal oise) chael capacity C = B log 2 (1 + S/N) where C is the maximum supportable bit rate B is the chael badwidth S/N is the ratio betwee sigal power ad i-bad oise power This N is oise ot umber of symbols Sprig 2018 CS 438 Staff, Uiversity of Illiois 64

How does Noise affect these Bouds? Noise Blurs the symbols, reducig the umber of symbols that ca be reliably distiguished Claude Shao (1948) Exteded Nyquist s work to chaels with additive white Gaussia oise (a good model for thermal oise) chael capacity C = B log 2 (1 + S/N) Represets error free capacity also used to calculate the oise that ca be tolerated to achieve a certai rate through a chael Result is based o may assumptios Formula assumes white oise (thermal oise) Impulse oise is ot accouted for Various types of distortio are also ot accouted for Sprig 2018 CS 438 Staff, Uiversity of Illiois 65

Noisy Capacity Telephoe chael 3400 Hz at 40 db SNR decibels (db) is a logarithmic uit of measuremet that expresses the magitude of a physical quatity (usually power or itesity) relative to a specified or implied referece level Sprig 2018 CS 438 Staff, Uiversity of Illiois 66

Decibels A ratio betwee sigal powers is expressed i decibels decibels (db) = 10log 10 (P 1 / P 2 ) Used i may cotexts The loss of a wireless chael The gai of a amplifier Note that db is a relative value Ca be made absolute by pickig a referece poit Decibel-Watt power relative to 1W Decibel-milliwatt power relative to 1 milliwatt Sprig 2018 CS 438 Staff, Uiversity of Illiois 67

Sigal-to-Noise Ratio Sigal-to-oise ratio (SNR, or S/N) Ratio of the power i a sigal to the power cotaied i the oise Typically measured at a receiver ( SNR ) = db 10log10 sigal power oise power A high SNR High-quality sigal Low SNR May be hard to extract the sigal from the oise SNR sets upper boud o achievable data rate Sprig 2018 CS 438 Staff, Uiversity of Illiois 68

Noisy Capacity Telephoe chael 3400 Hz at 40 db SNR C = B log 2 (1+S/N) bits/s SNR = 40 db 40 =10 log 10 (S/N) S/N =10,000 C = 3400 log 2 (10001) = 44.8 kbps Sprig 2018 CS 438 Staff, Uiversity of Illiois 69

More examples of Nyquist ad Shao Formulas Spectrum of a chael betwee 3 MHz ad 4 MHz ; SNR db = 24 db B = SNR = Usig Shao s formula C = B log 2 (1 + S/N) Sprig 2018 CS 438 Staff, Uiversity of Illiois 70

More examples of Nyquist ad Shao Formulas Spectrum of a chael betwee 3 MHz ad 4 MHz ; SNR db = 24 db B = 4 MHz - 3 MHz = 1MHz SNR SNR db = = 24 db = 10log 251 Usig Shao s formula C = 10 C = B log 2 (1 + S/N) 6 log 2 10 ( SNR) ( ) 6 1+ 251» 10 8 = 8Mbps Sprig 2018 CS 438 Staff, Uiversity of Illiois 71

More examples of Nyquist ad Shao Formulas How may sigalig levels are required? C = 2B log2m Sprig 2018 CS 438 Staff, Uiversity of Illiois 72

More examples of Nyquist ad Shao Formulas How may sigalig levels are required? C 8 10 4 = M = 2B 6 log 2 = 16 log = 2 2 M M ( 6 10 ) log 2 M Look out for: db versus liear values, log 2 versus log 10 Sprig 2018 CS 438 Staff, Uiversity of Illiois 73

Summary of Ecodig Problems: atteuatio, dispersio, oise Digital trasmissio allows periodic regeeratio Variety of biary voltage ecodigs High frequecy compoets limit to short rage More voltage levels provide higher data rate Modulatio schemes Amplitude, frequecy, phase, ad combiatios Quadrature amplitude modulatio: amplitude ad phase, may sigals Nyquist (oiseless) ad Shao (oisy) limits o data rates Sprig 2018 CS 438 Staff, Uiversity of Illiois 74