Mobile and Wireless Networking Ibrahim Korpeoglu

Mobile and Wireless Networking Ibrahim Korpeoglu
Introduction Mobile and Wireless Networking Ibrahim Korpeoglu CS 515 Ibrahim Korpeoglu

Outline Course Info Introduction History of Wireless
What is Wireless What is PCS History of Wireless Some Mobile Statistics CS 515 Ibrahim Korpeoglu

Course Information CS 515 Ibrahim Korpeoglu

Course Details Instructor: Ibrahim Korpeoglu Class Hours: Wed 15:40-16:30 Fri: 13:40-15:30 Office Hours Thu 10:40-12:00 Classroom: EA 502 (You can also come to my office at any time if you need to see me) CS 515 Ibrahim Korpeoglu

Recommended Textbooks
Theodore Rappaport, Wireless Communications: Principles and Practice, Second Edition, Prentice Hall, December 2001. Yi-Bing Lin, Imrich Chlamtac, Wireless and Mobile Network Architectures, John Wiles & Sohns, 1st edition, 2000. You don’t have to buy these books. But I recommend buying them if you have the opportunity! CS 515 Ibrahim Korpeoglu

Reading List You will read a lot of papers in this course
The papers are on the course web page You can download them from there. If a paper is not there, let me know. I will put the paper on my door if there is no online copy of the paper The paper-list size on the webpage will be reduced, so that you don’t spend all of your time only on this course. CS 515 Ibrahim Korpeoglu

Grading There will be one midterm and one final exam
There may be projects. I did not determine them yet. Simulation or implementation projects No idea how hard they will be! No idea which language(s) they will be implemented on! Attendance is important! CS 515 Ibrahim Korpeoglu

Why projects are important?
I hear and I forget, I see and I remember, I do and I understand Confucius CS 515 Ibrahim Korpeoglu

Outline Introduction What is wireless and mobile networking
History of Wireless Challenges of Mobile and Wireless Communication and Networking What is Personal Communications Systems Why there is demand on that What is ubiquitous computing. Overview of Wireless Technologies and Systems CS 515 Ibrahim Korpeoglu

Outline Wireless Link Characteristics Radio Propagation
Short and Long wave properties Attenuation Interfence Fading and Multi-path Fading Transmit power and range Bit Error Rate and Models CS 515 Ibrahim Korpeoglu

Outline Wireless Media Access
What is different in Wireless Media than Wireline Media Why CSMA/CD does not work MACA and MACAW protocols TDMA and FDMA CDMA CS 515 Ibrahim Korpeoglu

Outline Handoff Routing More from telecom point view
How handoffs are triggered How handoffs are managed Routing more from data networking point of view How mobility affect routing for mobile hosts Mobile IP CS 515 Ibrahim Korpeoglu

Outline Transport Protocols over Wireless and Mobile Networks
How does wireless links and mobile hosts affect the performance and operation of transport protocols Look specifically to TCP There are many proposals to improve the performance of TCP over wireless links and for mobile hosts CS 515 Ibrahim Korpeoglu

Outline Ad-Hoc Mobile Networks
What if the mobile hosts are not roaming around an infrastructure-based network Ad-hoc networks are established spontaneounly There is no infrastructure that you can rely on A mobile terminal may also act as an network router Routing protocols for ad-hoc networks Network connectivity graph is not fixed; dynamically changes over time The network elements are small-capacity, battery-powered devices CS 515 Ibrahim Korpeoglu

Outline Looking closely to the wireless systems
Wireless Local Area Networks and HiperLAN Standards Wireless Personal Area Networks and Home Networking Bluetooth and HomeRF Wide-Area Wireless Cellular Networks GSM CDMA GPRS 3G Networks CS 515 Ibrahim Korpeoglu

Outline Wireless and Mobile Applications
Wireless Application Protocol Mobile Applications Mobile Databases Quality of Service in Mobile/Wireless Networks What are the challenges for providing QoS in mobile and wireless environments CS 515 Ibrahim Korpeoglu

Outline Service and Device Discover in Mobile Networks
How can you discover the resources around you Service Location Protocol Jini Power Management How low-power objective affect the design of wireless systems and network protocols Issues and solutions CS 515 Ibrahim Korpeoglu

Outline Introduction to Peer2peer networking Wrap up and Conclusions
What is peer2peer networking Why client-server computing is not enough always Centralized, distributed and hybrid peer2peer systems Wrap up and Conclusions CS 515 Ibrahim Korpeoglu

What is Wireless and Mobile Communication?
CS 515 Ibrahim Korpeoglu

Wireless Communication
Transmitting voice and data using electromagnetic waves in open space Electromagnetic waves Travel at speed of light (c = 3x108 m/s) Has a frequency (f) and wavelength (l) c = f x l Higher frequency means higher energy photons The higher the energy photon the more penetrating is the radiation CS 515 Ibrahim Korpeoglu

Electromagnetic Spectrum
104 102 100 10-2 10-4 10-6 10-8 10-10 10-12 10-14 10-16 Radio Spectrum Micro wave IR UV X-Rays Cosmic Rays 104 106 108 1010 1012 1014 1016 1018 1020 1022 1024 1MHz ==100m 100MHz ==1m 10GHz ==1cm Visible light < 30 KHz VLF KHz LF 300KHz – 3MHz MF 3 MHz – 30MHz HF 30MHz – 300MHz VHF 300 MHz – 3GHz UHF 3-30GHz SHF > 30 GHz EHF CS 515 Ibrahim Korpeoglu

Wavelength of Some Technologies
GSM Phones: frequency ~= 900 Mhz wavelength ~= 33cm PCS Phones frequency ~= 1.8 Ghz wavelength ~= 17.5 cm Bluetooth: frequency ~= 2.4Gz wavelength ~= 12.5cm CS 515 Ibrahim Korpeoglu

Frequency Carries/Channels
The information from sender to receiver is carrier over a well defined frequency band. This is called a channel Each channel has a fixed frequency bandwidth (in KHz) and Capacity (bit-rate) Different frequency bands (channels) can be used to transmit information in parallel and independently. CS 515 Ibrahim Korpeoglu

Example Assume a spectrum of 90KHz is allocated over a base frequency b for communication between stations A and B Assume each channel occupies 30KHz. There are 3 channels Each channel is simplex (Transmission occurs in one way) For full duplex communication: Use two different channels (front and reverse channels) Use time division in a channel Station A Channel 1 (b - b+30) Station B Channel 2 (b+30 - b+60) Channel 3 (b+60 - b+90) CS 515 Ibrahim Korpeoglu

Homework 1 Read and digest the following papers!
M. Weiser, The Computer for the Twenty-First Century, Scientific American, Vol. 265, No. 3, (September 1991), pp D. Cox, Wireless Personal Communications: What is It?, IEEE Personal Communications Magazine, (April 1995), pp These papers are on the course webpage! CS 515 Ibrahim Korpeoglu

Simplex Communication
Normally, on a channel, a station can transmit only in one way. This is called simplex transmision To enable two-way communication (called full-duplex communication) We can use Frequency Division Multiplexing We can use Time Division Multiplexing CS 515 Ibrahim Korpeoglu

Duplex Communication - FDD
FDD: Frequency Division Duplex Base Station B Mobile Terminal M Forward Channel Reverse Channel Forward Channel and Reverse Channel use different frequency bands CS 515 Ibrahim Korpeoglu

Duplex Communication - TDD
TDD: Time Division Duplex Base Station B Mobile Terminal M M B M B M B A singe frequency channel is used. The channel is divided into time slots. Mobile station and base station transmits on the time slots alternately. CS 515 Ibrahim Korpeoglu

Example - Frequency Spectrum Allocation in U.S. Cellular Radio Service
Reverse Channel Forward Channel 991 992 … 1023 1 2 … 799 991 992 … 1023 1 2 … 799 MHz MHz Channel Number Center Frequency (MHz) Reverse Channel 1 <=N <= 799 991 <= N <= 1023 0.030N 0.030(N-1023) Forward Channel 1 <=N <= 799 991 <= N <= 1023 0.030N 0.030(N-1023) (Channels are unused) Channel bandwidth is 45 MHz CS 515 Ibrahim Korpeoglu

What is Mobility Initially Internet and Telephone Networks is designed assuming the user terminals are static No change of location during a call/connection A user terminals accesses the network always from a fixed location Mobility and portability Portability means changing point of attachment to the network offline Mobility means changing point of attachment to the network online CS 515 Ibrahim Korpeoglu

Degrees of Mobility Walking Users Vehicles Low speed
Small roaming area Usually uses high-bandwith/low-latency access Vehicles High speeds Large roaming area Usually uses low-bandwidth/high-latency access Uses sophisticated terminal equipment (cell phones) CS 515 Ibrahim Korpeoglu

The Need for Wireless/Mobile Networking
Demand for Ubiquitous Computing Anywhere, anytime computing and communication You don’t have to go to the lab to check your Pushing the computers more into background Focus on the task and life, not on the computer Use computers seamlessly to help you and to make your life more easier. Computers should be location aware Adapt to the current location, discover services CS 515 Ibrahim Korpeoglu

Some Example Applications of Ubiquitous Computing
You walk into your office and your computer automatically authenticates you through your active badge and logs you into the Unix system You go to a foreign building and your PDA automatically discovers the closest public printer where you can print your schedule and give to your friend CS 515 Ibrahim Korpeoglu

More Examples You walk into a Conference room or a shopping Mall with your PDA and your PDA is smart enough to collect and filter the public profiles of other people that are passing nearby Of course other people should also have smart PDAs. The cows in a village are equipped with GPS and GPRS devices and they are monitored from a central location on a digital map. No need for a person to guide and feed them You can find countless examples CS 515 Ibrahim Korpeoglu

How to realize Ubiquitous Computing
Small and different size computing and communication devices Tabs, pads, boards PDAs, Handhelds, Laptops, Cell-phones A communication network to support this Anywhere, anytime access Seamless, wireless and mobile access Need for Personal Communication Services (PCS) Ubiquitous Applications New software CS 515 Ibrahim Korpeoglu

What is PCS Personal Communication Services
Ibrahim Korpeoglu

What is PCS Personal Communication Services
A wide variety of network services that includes wireless access and personal mobility services Provided through a small terminal Enables communication at any time, at any place, and in any form. The market for such services is tremendously big Think of cell-phone market CS 515 Ibrahim Korpeoglu

Several PCS systems High-tier Systems
GSM: Global System for Mobile Communications The mobile telephony system that we are using IS-136 USA digital cellular mobile telephony system TDMA based multiple access Personal Digital Cellular IS-95 cdmaOne System CDMA based multiple access CS 515 Ibrahim Korpeoglu

Several PCS systems Low-tier systems
Residential, business and public cordless access applications and systems Cordless Telephone 2 (CT2) Digital Enhanced Cordless Telephone (DECT) Personal Access Communication Systems (PACS) Personal Handy Telephone System (PHS) CS 515 Ibrahim Korpeoglu

Several PCS systems Wideband wireless systems
For Internet access and multimedia transfer Cdma2000 W-CDMA, proposed by Europe SCDMA, proposed by Chine/Europe CS 515 Ibrahim Korpeoglu

Several PCS systems Other PCS Systems Special data systems
CDPD: Cellular Digital Packet Data RAM Mobile Data Advanced Radio Data Information System (ARDIS) Paging Systems Mobile Satellite Systems LEO, MEO, HEO satellites for data/voice ISM band systems: Bluetooth, , etc. CS 515 Ibrahim Korpeoglu

PCS Problems How to integrate mobile and wireless users to the Public Switched Telephone Network (PSTN) (Voice Network) Cellular mobile telephony system How to integrate mobile and wireless users to the Internet (Data Network) Mobile IP, DHCP, Cellular IP How to integrate all of them together and also add multimedia services (3G Systems) CS 515 Ibrahim Korpeoglu

Looking to PCS from different Angles
Internet PSTN (Telephone Network) Wireless Access Mobile Users Laptop users Pocket PC users Mobile IP, DHCP enabled computers Mobile Users Cell phone users Cordless phone users Telecom People View Data Networking People View CS 515 Ibrahim Korpeoglu

What does this course cover?
This course will cover the problems/solutions in the telecommunication domain and also in the data networking domain Mobile IP (data) TCP over Wireless (data) GSM, GPRS, CDMA (telecom) We will also cover some fundamental problems/solutions for wireless access Wireless channel characteristics Recovering from errors Wireless media access CS 515 Ibrahim Korpeoglu

Telecom and Data Networking
Telecom Interest Data Networking Interest - Voice Transmission - Frequency Reuse Handoff Management Location Tracking Roaming QoS GSM, CDMA, Cordless Phones, GPRS, EDGE Data Transmission Mobile IP (integrating mobile hosts to internet) Ad-hoc Networks TCP over Wireless Service Discovery Radio Propagation Link Characteristics Error Models -Wireless Medium Access (MAC) - Error Control CS 515 Ibrahim Korpeoglu

Very Basic Cellular/PCS Architecture
Mobility Database Public Switched Telephone Network Base Station Controller Mobile Switching Center (MSC) Radio Network Base Station (BS) Mobile Station CS 515 Ibrahim Korpeoglu

Wireless System Definitions
Mobile Station A station in the cellular radio service intended for use while in motion at unspecified locations. They can be either hand-held personal units (portables) or installed on vehicles (mobiles) Base station A fixed station in a mobile radio system used for radio communication with the mobile stations. Base stations are located at the center or edge of a coverage region. They consists of radio channels and transmitter and receiver antennas mounted on top of a tower. CS 515 Ibrahim Korpeoglu

Mobile Switching Center Switching center which coordinates the routing of calls in a large service area. In a cellular radio system, the MSC connections the cellular base stations and the mobiles to the PSTN (telephone network). It is also called Mobile Telephone Switching Office (MTSO) Subscriber A user who pays subscription charges for using a mobile communication system Transceiver A device capable of simultaneously transmitting and receiving radio signals CS 515 Ibrahim Korpeoglu

Control Channel Radio channel used for transmission of call setup, call request, call initiation and other beacon and control purposes. Forward Channel Radio channel used for transmission of information from the base station to the mobile Reverse Channel Radio channel used for transmission of information from mobile to base station CS 515 Ibrahim Korpeoglu

Simplex Systems Communication systems which provide only one-way communication Half Duplex Systems Communication Systems which allow two-way communication by using the same radio channel for both transmission and reception. At any given time, the user can either transmit or receive information. Full Duplex Systems Communication systems which allow simultaneous two-way communication. Transmission and reception is typically on two different channels (FDD). CS 515 Ibrahim Korpeoglu

Handoff The process of transferring a mobile station from one channel or base station to an other. Roamer A mobile station which operates in a service area (market) other than that from which service has been subscribed. Page A brief message which is broadcast over the entire service area, usually in simulcast fashion by many base stations at the same time. CS 515 Ibrahim Korpeoglu

PCS Systems Classification
Cordless Telephones Cellular Telephony (High-tier) Wide Area Wireless Data Systems (High-tier) High Speed Local and Personal Area Networks Paging Messaging Systems Satellite Based Mobile Systems 3G Systems CS 515 Ibrahim Korpeoglu

Major Mobile Radio Standards USA
Type Year Intro Multiple Access Frequency Band (MHz) Modulation Channel BW (KHz) AMPS Cellular 1983 FDMA FM 30 USDC 1991 TDMA DQPSK CDPD 1993 FH/Packet GMSK IS-95 Cellular/PCS CDMA QPSK/BPSK 1250 FLEX Paging Simplex Several 4-FSK 15 DCS-1900 (GSM) PCS 1994 200 PACS Cordless/PCS TDMA/FDMA 300 CS 515 Ibrahim Korpeoglu

Major Mobile Radio Standards - Europe
Type Year Intro Multiple Access Frequency Band (MHz) Modulation Channel BW (KHz) ETACS Cellular 1985 FDMA 900 FM 25 NMT-900 1986 12.5 GSM Cellular/PCS 1990 TDMA GMSK 200KHz C-450 20-10 ERMES Paging 1993 FDMA4 Several 4-FSK CT2 Cordless 1989 GFSK 100 DECT 1728 DCS-1800 Cordless/PCS 200 CS 515 Ibrahim Korpeoglu

Cordless Telephones PSTN Telephone Network Base unit Cordless Phone

Cordless Telephones Characterized by Appeared as analog devices
Low mobility (in terms of range and speed) Low power consumption Two-way tetherless (wireless) voice communication High circuit quality Low cost equipment, small form factor and long talk-time No handoffs between base units Appeared as analog devices Digital devices appeared later with CT2, DECT standards in Europe and ISM band technologies in USA CS 515 Ibrahim Korpeoglu

Cordless Telephones Usage Design Choices At homes
At public places where cordless phone base units are available Design Choices Few users per MHz Few users per base unit Many base units are connected to only one handset Large number of base units per usage area Short transmission range CS 515 Ibrahim Korpeoglu

Cordless Phone Some more features
32 Kb/s adaptive differential pulse code modulation (ADPCM) digital speech encoding Tx power <= 10 mW Low-complexity radio signal processing No forward error correction (FEC) or whatsoever. Low transmission delay < 50ms Simple Frequency Shift Modulation (FSK) Time Division Duplex (TDD) CS 515 Ibrahim Korpeoglu

Cellular Telephony Characterized by High mobility provision Wide-range
Two-way tetherless voice communication Handoff and roaming support Integrated with sophisticated public switched telephone network (PSTN) High transmit power requires at the handsets (~2W) CS 515 Ibrahim Korpeoglu

Cellular Telephony - Architecture

Cellular Telephony Systems
Mobile users and handsets Very complex circuitry and design Base stations Provides gateway functionality between wireless and wireline links ~1 million dollar Mobile switching centers Connect cellular system to the terrestrial telephone network CS 515 Ibrahim Korpeoglu

World Cellular Subscriber Growth

Mobile Systems Market Ericsson sells half of the mobile base stations
1 base station ~ 100 thousand - 1 million dollar Nokia has the biggest market in cell-phones 1 cell-phone ~ 100 dollar Nokia has to sell 10,000 cell-phones to match the revenue Ericsson obtains from selling just one base-station! CS 515 Ibrahim Korpeoglu

Cellular Networks First Generation Second Generation (2G) 2.5G
Analog Systems Analog Modulation, mostly FM AMPS Voice Traffic FDMA/FDD multiple access Second Generation (2G) Digital Systems Digital Modulation TDMA/FDD and CDMA/FDD multiple access 2.5G Voice + Low-datarate Data Third Generation Digital Voice + High-datarate Data Multimedia Transmission also CS 515 Ibrahim Korpeoglu

2G Technologies cdmaOne (IS-95) GSM, DCS-1900 IS-54/IS-136 PDC
Uplink Frequencies (MHz) (Cellular) (US PCS) MHz (Eurpe) (US PCS) 800 MHz, 1500 Mhz (Japan) Downlink Frequencies MHz (US Cellular) MHz (US PCS) (Europa) (US PCS) MHz (Cellular) (US PCS) 800 MHz, 1500 MHz (Japan) Deplexing FDD Multiple Access CDMA TDMA Modulation BPSK with Quadrature Spreading GMSK with BT=0.3 p/4 DQPSK Carrier Seperation 1.25 MHz 200 KHz 30 KHz (IS-136) (25 KHz PDC) Channel Data Rate Mchips/sec Kbps 48.6 Kbps (IS-136) 42 Kbps (PDC) Voice Channels per carrier 64 8 3 Speech Coding CELP at 13Kbps EVRC at 8Kbps RPE-LTP at 13 Kbps VSELP at 7.95 Kbps CS 515 Ibrahim Korpeoglu

2G and Data 2G is developed for voice communications
You can send data over 2G channels by using modem Provides adat rates in the order of ~9.6 Kbps Increased data rates are requires for internet application This requires evolution towards new systems: 2.5 G CS 515 Ibrahim Korpeoglu

2.5 Technologies Evolution of TDMA Systems Evolution of CDMA Systems
HSCSD for 2.5G GSM Up to 57.6 Kbps data-rate GPRS for GSM and IS-136 Up to Kbps data-rate EDGE for 2.5G GSM and IS-136 Up to 384 Kbps data-rate Evolution of CDMA Systems IS-95B Up to 64 Kbps CS 515 Ibrahim Korpeoglu

3G Systems Goals Voice and Data Transmission
Simultanous voice and data access Multi-megabit Internet access Interactive web sessions Voice-activated calls Multimedia Content Live music CS 515 Ibrahim Korpeoglu

3G Systems Evolution of Systems CDMA sysystem evaolved to CDMA2000
CDMA2000-1xRTT: Upto 307 Kbps CDMA2000-1xEV: CDMA2000-1xEVDO: upto 2.4 Mbps CDMA2000-1xEVDV: 144 Kbps datarate GSM, IS-136 and PDC evolved to W-CDMA (Wideband CDMA) (also called UMTS) Up to Mbps data-rates Future systems 8Mbps Expected to be fully deployed by New spectrum is allocated for these technologies CS 515 Ibrahim Korpeoglu

Interest to 3G Applications
Western Eastern USA Europe Europe s City maps/directions Latest news Authorize/enable payment Banking/trading online Downloading music Shopping/reservation Animated images Chat rooms, forums Interactive games Games for money (Means based upon a six-point interest scale, where 6 indicates high interest and 1 indicates low interest.) CS 515 Ibrahim Korpeoglu

Upgrade Paths for 2G Technologies
IS-95 GSM IS-136 PDC 2.5G GPRS HSCSD IS-95B EDGE 3G cdma200-1xRTT W-CDMA EDGE cdma2000-1xEV,DV,DO TD-SCDMA cdma200-3xRTT CS 515 Ibrahim Korpeoglu

GSM Subscriber Growth CS 515 Ibrahim Korpeoglu

CDMA Subscriber Growth

CDMA2000 Subscriber Growth

GSM and CDMA Coverage Map Worldwide

GSM Networks in Turkey Number of Subscribers (Nov 2001)
Network System GPRS HSCSD Frequency Aria GSM Live (March 2002) no Aycell GSM no no Telsim GSM Live (Aug. 2000) no Turkcell GSM Live (March 2001) soon Number of Subscribers (Nov 2001) Turkcell: 6,800,900 Telsim: 2,800,000 CS 515 Ibrahim Korpeoglu

Coverage Map - Turkcell

Coverage Map - Telsim CS 515 Ibrahim Korpeoglu

Coverage Map - Aria CS 515 Ibrahim Korpeoglu

Coverage Map - Aycell CS 515 Ibrahim Korpeoglu

Mobile Phone Market Share
1st Quarter of 2002 Nokia % Motorola % Samsung % Siemens % Sony-Ericsson 6.4% CS 515 Ibrahim Korpeoglu

Some Mobile Statistics – June 2002
Total Global Mobile Users: 860m Total Analog Users: 71m Total US Mobile Users: 137.5m Total GSM Users: 669m Total TDMA Users: 84m Total European Users: 279m Global Montly SMSs/User: 36 SMS Sent in 2001: billion GSM Countries on Air: 171 #1 Mobile Country: China #1 GSM Country: China #1 SMS Country: Phillipines #1 Cell Phone Vendor: Nokia #1 Network in Europa: T-Mobil #1 Network in Japan: DoCoMo #1 Telecom Infrastructure Company: Ericsson CS 515 Ibrahim Korpeoglu

Introduction to Wireless Systems and History
CS 515 Mobile and Wireless Networking İbrahim Korpeoğlu Computer Engineering Department Bilkent University, Ankara CS 515 Ibrahim Korpeoglu

Objectives and Compromises in Designing Cellular Systems
Maximize users per MHz Maximize users per cell cite Base stations are too costly Compromises made for these objectives High power transmitters at base stations to increase the range High power transmitters for user-sets High user-set complexity High power consumption due to complex digital signal processing High network complexity CS 515 Ibrahim Korpeoglu

Features of Cellular Systems
Low bit-rate speech coding: <= 13 Kb/s Increases the number of users per MHz and cell-site Decreases the voice quality Some systems (like CDMA) make use of speech inactivity High Transmission Delay 200 ms round-trip time High-complexity DSP (digital dignal processing) For speech coding and de-modulatioin CS 515 Ibrahim Korpeoglu

Features of Cellular Systems
Fixed Channel Allocation Channel are allocated to cell-sites statically Dynamic allocation is complex to work also with handoffs Frequency Division Duplex Network and system complexity for providing synchronization between network elements is relieved TDD requires synchronized clock at both ends Mobile handset power control Decreases the co-channel interference Hence increases the system capacity Implemented in CDMA systems CS 515 Ibrahim Korpeoglu

System Capacity System Capacity is the number all users that can communicate (use the system) at the same time A base station (cell) has a fixed number of channels available, hence at a given time a fixed number of users can talk simultaneously CS 515 Ibrahim Korpeoglu

System Capacity(C) and Coverage Area(d2)
Low cost base-stations covering a small area High cost base-stations covering a large area Each base station has a fixed number of channels for both systems All channels in System1 = 9 x All channels in System2 C1 / C2 = (d2 / d1)2 CS 515 Ibrahim Korpeoglu

System Capacity/Coverage Area/Economics
Compare high-tier and low-tier PCS systems Assume channels/MHz are the same for both systems High-tier System Cell-site spacing: 20,000 ft (6km), cost = $ 1M Low-tier System Cell-site spacing: 1,000 ft (300m) Low-tier System Capacity = (20000/1000)2 = 400xHigh-tier Capacity Then, for having both system costs the same Base station costs in low-tier systems should be $ 1M / 400 = $ 2500 CS 515 Ibrahim Korpeoglu

Wide Area Data Systems Characterized by
Providing high mobility Wide-range Low data-rate 9.2 Kbps, 19.2 Kbps Digital data communication Deployed in several cities and regions in USA Some Systems like CDPD are overlayed over existing cellular system (AMPS) CDPD: Cellular Digital Packet Data CDPD uses the unused voice channels in AMPS system for data transmission CS 515 Ibrahim Korpeoglu

Wide Area Wireless Packet Data Systems
CDPD RAM Mobile (Mobitex) ARDIS Metricom Data Rate 19.2 Kbps 8 Kbps 4.8 Kbps ~176 Kbps Modulation GMSK Frequency ~800MHz ~900MHz ~915MHz Channel Spacing 30 KHz 12.5KHz 25KHz 160KHz Access Means Unused AMPS Channels Slotted Aloha CSMA FHSS (ISM) Transmit Power 40 watt 1 watt CS 515 Ibrahim Korpeoglu

Example: Metricom Provides wireless Internet access for desktop, laptop and palmtop computers A Metricom modem can connect to the serial/USB port or can connect as a PCMCIA card to the main bus. A metricom modem has around 1 mile range Uses a Microcellular Packet Radio Network Frequency band: ISM band at MHz (915 MHz middle freq.) Power control is used to minimize interference and maximize battery lifetime Systems Shoe-box sizes base station Mobile terminal wireless adapters Deployed in Bay Area, San Francisco. CS 515 Ibrahim Korpeoglu

Metricom - How it works Computers use Metricom modems that send data via radio signals to base stations located on utility poles The pole-top radios send the data via radio signal to wired antennas The radio signals are converted to wire-based signals at the antenna and sent to the metricom network interface over high-speed network Metricom network interface connects directly to the internet and private networks CS 515 Ibrahim Korpeoglu

Metricom Coverage - Bay Area, San Francisco

High-Speed Wireless LANs (WLAN)
Characterized by Low mobility (not for vehicular use) High speed data transmission Confined regions – buildings and campuses Coverage: 100m – 300m per base station Speed: 2-11Mbps, 20Mbps Uses ISM bands MHz MHz MHz Uses FHSS or DSSS spectrum usage techniques CS 515 Ibrahim Korpeoglu

WLAN Standards Bitrate Frequency Band Range IEEE 802.11b 5.5 – 11Mbps
2.4 GHz ~100m IEEE a 54 Mbps 5 Ghz HiperLAN (Europe) 20Mbps 5 GHz ~50m HiperLAN/2 CS 515 Ibrahim Korpeoglu

Personal Area Networks (PANs)
Bluetooth 2.5GHz ISM band 10m range, 1mW transmit power 100m range, requires increase in transmit power 1 Mbps data rate shared between 7 devices FHSS spread spectrum use TDD duplex scheme Polling based multiple access Retricted start topology 1 master connects to 7 slaves CS 515 Ibrahim Korpeoglu

Paging Systems Categorized as
One-way messaging (asymmetric communication) Wide-area coverage Low complexity, very low-power pager (receiver) devices CS 515 Ibrahim Korpeoglu

Wide-Area Paging System
City 1 Terrestrial Link Paging Terminal City 2 Paging Terminal Terrestrial Link Paging Control Center City N Paging Terminal Satellite Link CS 515 Ibrahim Korpeoglu

Satellite Based Mobile Systems
Categorized as Two-way (or one-way) limited quality voice or data transmission Very wide range and coverage Large regions Sometimes global coverage Very useful in sparsely populated areas: rural areas, sea, mountains, etc. Target: Vehicles and/or other stationary/mobile uses Expensive base station (satellites) systems CS 515 Ibrahim Korpeoglu

Satellite based systems
Very large coverage Low overall system capacity Expensive service Proposed Satellite Systems LEOS: Low-earth orbit satellite systems satellites/system High overall system capacity, low delay Many but comparably less expensive satellites MEOS: Medium-earth satellite systems GEOS: Geostationary or Geosynchronous Orbit Systems Fewer than 10 satellites/system Low overall system capacity, high end-to-end delay (~0.5sec) Very expensive satellites Iridium, Globalstar, Teledesic, Inmarsat are some example systems CS 515 Ibrahim Korpeoglu

History of Wireless and Mobile Communication

History 1831: Faraday had first started experimenting with electromagnetic waves. Electromagnetic wave: one of the waves that are propagated by simultaneous periodic variations of electric and magnetic field intensity and that include radio waves infrared visible light ultraviolet, X rays Gamma rays CS 515 Ibrahim Korpeoglu

History – Mathematics and EM
1864: Maxwell who had been working on a mathematical model for electromagnetic waves finally published his paper on the subject. One of the consequences of his theories was that E.M. waves would travel at near the speed of light. This had also been experimentally determined by others at the time. CS 515 Ibrahim Korpeoglu

History At the time very few people fully understood Maxwell’s equation. Three scientists who have been called Maxwellians studied his equations further: Lodge, Hertz, Tesla. Oliver Lodge realized that the equations not only implied the nature of light and heat, but also that there was a whole spectrum of radiation with wavelengths both above and below those of visible light. He was probably one of the first to realize that this could be produced electrically. CS 515 Ibrahim Korpeoglu

History – Existence of EM Waves
At the same time that Lodge was carrying out his experiments, Heinrich Hertz in Germany was also doing some of his own concerning Maxwell’s equations. Hertz's investigations into Maxwell’s equations involved generation, detection, and measurement of waves in free space, rather than along wires. 1887: Hertz proves existence of EM waves; first spark transmitter generates a spark in a receiver several meters away The units of frequency waves is named after him, 1 cycle/second equals a Hertz. CS 515 Ibrahim Korpeoglu

History – Wireless Telegraph
1896: Guglielmo Marconi demonstrates wireless telegraph to English telegraph office 1897: ``The Birth of Radio'' - Marconi awarded patent for wireless telegraph 1897: First ``Marconi station'' established on Needles island to communicate with English coast 1898: Marconi awarded English patent no for tuned communication 1898: Wireless telegraphic connection between England and France established CS 515 Ibrahim Korpeoglu

History – Frequency Tuning
In 1898: Tesla gave one of the first wireless demonstrations with a what we would call a remote control boat. He realized that things of this nature would need to only respond to their own frequency, and remain inactive otherwise. This was Tesla’s fundamental radio tuning invention, which he had first described several years earlier. CS 515 Ibrahim Korpeoglu

History – Transoceanic Communication
1901: Marconi successfully transmits radio signal across Atlantic Ocean from Cornwall to Newfoundland 1902: First bidirectional communication across Atlantic 1909: Marconi awarded Nobel prize for physics CS 515 Ibrahim Korpeoglu

History – Voice over Radio
1914: First voice over radio transmission 1920s: Mobile receivers installed in police cars in Detroit 1930s: Mobile transmitters developed; radio equipment occupied most of police car trunk by 1934: Amplitude Modulation (AM) systems used by police cars and stations 1935: Edwin Armstrong demonstrated frequency modulation (FM) for the first time. Majority of police systems converted to FM CS 515 Ibrahim Korpeoglu

History – Mobile Telephony
1946: First public mobile telephone service was introduced. First interconnection of mobile users to public switched telephone network (PSTN) 1949: FCC (Federal Communications Commission) of US recognizes mobile radio as new class of service : AT&T Bell Labs developed theory and techniques for cellular telephony CS 515 Ibrahim Korpeoglu

History - Pager 1959: The term "pager" was first used, referring to a Motorola radio communications product 1968: AT&T proposed cellular telephony to FCC of US. 1974: The first pager was introduced by Motorola. 1977: Public cell phone testing began. 1979: World’s first cellular system was implemented by NTT Japan. 1980: 3.2 million pagers used wordwide. They had limited range. CS 515 Ibrahim Korpeoglu

History – Cordless Phones and Cellular Telephony
1980: Cordless phones started to emerge. Early 1980s: Wireless modems emerged. 1981: European Nordic Mobile Telephone (NMT) System was developed 1983: FCC allocated wireless spectrum for mobile telephony. 1983: AMPS, first USA analog cellular telephony standard was developed CS 515 Ibrahim Korpeoglu

History – Wireless Data
1983: Intoduction of ARDIS wireless data service 1985: European Total Access Cellular System (ETACS) was deployed. 1985: In Germany, cellular standard C-450 was introduced. 1985: ISM bands defined for commercial spread spectrum applications CS 515 Ibrahim Korpeoglu

History – Wireless LANs, GSM
1987: IEEE Wireless LAN working grup founded. 1989: In Europe, GSM was defined. 1990: In Europe, GSM deployed. by 1990: Wide-area paging had been invented and over 22 million pagers were in use 1990: FCC allocated spctrum in 900 Mhz for cordless phones. 1990: Announcement of Wireless LAN products CS 515 Ibrahim Korpeoglu

History 1991: First US digital cellular hardware was installed. IS-54 and IS-136 emerged. 1991: RAM mobile (mobitex) data service 1992: HyperLAN in Europa 1992: World Radio Conference in Malaga (WRC- 92) allocated frequencies for future UMTS use. Frequencies and MHz were identified for IMT2000 use CS 515 Ibrahim Korpeoglu

History 1993: First GSM 1800 system in commercial operation in UK
1993: IS-95 code-division multiple-access (CDMA) spread- spectrum digital cellular system deployed in US 1993: CDPD (Cellular Digital Packet Data) over AMPS was realized 1994: GSM system deployed in US 1994: there were over 61 million pagers in use and pagers became popular for personal use. CS 515 Ibrahim Korpeoglu

History – Bluetooth, PCS
1994: Ericsson starts investigating a low-power, low-cost radio technology to remove cables around cell phones (born of Bluetooth idea) 1995: FCC auctions off frequencies in Personal Communications System (PCS) band at 1.8 GHz for mobile telephony 1995: DSS started to be used for cordless phones CS 515 Ibrahim Korpeoglu

History – Third Generation
1995 The UMTS Task Force was established 1996: The UMTS Forum was established in Zurich. 1997 the UMTS Forum produced its first report entitled 1997: IEEE has been standardized (2 Mbps) 1997: IS-95B standard complete; includes 64 kbps data CS 515 Ibrahim Korpeoglu

History – Personal Area Networks
1998: Bluetooth was born. SIG for Bluetooth has been established by the leadership of 5 companies: Ericsson, IBM, Intel, Toshiba, Nokia 1998: HomeRF Working Group was formed. 1998: FCC gave 2.5 GHz spectrum for cordless phones 1998 ETSI SMG meeting in Paris both W-CDMA and TD-CDMA proposals were combined to UMTS air interface specification. CS 515 Ibrahim Korpeoglu

History – 3G Trials and Progress
1998: The first call using a Nokia W-CDMA terminal in DoCoMo's trial network was completed at Nokia's R&D unit near Tokyo in Japan. Jun 1998: CDMA2000 submitted to ITU for IMT Dec 1998: The first meetings of the 3GPP Technical Specification Groups in France. 1999: IEEE b approved (11 Mbps) 1999: The first open Bluetooth specification 1.0 is released. CS 515 Ibrahim Korpeoglu

Histor – 3G Progress Jul 1999: Phase 1 CDMA2000 standard complete and approved for publication Jul 1999: Korea Telecom Freetel launches world's first IS-95B network in Korea 1999: Nokia Oyj said that it has completed what it claims to be the first WCDMA call through the public switched telephone network in the world Nov 1999: ITU-R Task Group 8/1 endorses CDMA2000 standards (three modes) for IMT CS 515 Ibrahim Korpeoglu

History – 3G Progress 1999: ETSI Standardization finished for UMTS Release 1999 specifications both for FDD and TDD in Nice, France. Mar 1999: March 1999 ITU approves radio interfaces for third generation mobile systems 1999: World Radio Conference (WRC-99) handled spectrum and regulatory issues for advanced mobile communications applications in the context of IMT-2000 June 2000: Telstra and Nortel complete first 3G CDMA2000 1X data transmission CS 515 Ibrahim Korpeoglu

History – Bluetooth on the Market
2000: The first certified Bluetooth products on the market Oct 2000: SK Telecom and LG Telecom (Korea) launch world's first 3G commercial services using CDM2000 Mar 2001: 3GPP approves UMTS Release 4 specification in Palm Springs, CA. 2001: The latest Bluetooth protocol 1.1 is released. CS 515 Ibrahim Korpeoglu

History – 3G Progress 2001 Ericsson and Vodafone UK claim to have made the world's first WCDMA voice call over commercial network. Jun 2001: NTT DoCoMo launched a trial 3G service June 2001: CDMA2000 1xEV-DO recognized as part of the 3G IMT-2000 standard CS 515 Ibrahim Korpeoglu

History – 3G Commercial Services
Aug 2001: 1 million commercial CDMA2000 1X subscribers Oct 2001 NTT DoCoMo launched the first commercial WCDMA 3G mobile network Nov 2001: Nokia and AT&T Wireless complete first live 3G EDGE call. Dec 2001: Telenor launched in Norway the first commercial UMTS network Jan 2002: Verizon Wireless (US) launches commercial CDMA2000 1X service CS 515 Ibrahim Korpeoglu

History – This Year and Future
Jan 2002: Verizon Wireless (US) launches commercial CDMA2000 1X service Feb 2002: Nokia and Omnitel Vodafone claims to have made the first rich call in an end-to-end All- IP mobile network at the 3GSM World Congress in Cannes, France. May 2002: 10 million commercial CDMA2000 1X subscribers Jun 2003: Target date for UMTS Release 6 2005: UMTS service will be world-wide CS 515 Ibrahim Korpeoglu

Challenges of Mobile and Wireless Computing
CS 515 Mobile and Wireless Networking İbrahim Körpeoğlu Computer Engineering Department Bilkent University, Ankara CS 515 Ibrahim Korpeoglu

Homework 2 Read and digest the following papers
H. Forman, J. Zahorjan, The Challenges of Mobile Computing, IEEE Computer, V 27, N 4, (April 1994), pp T. La Porta et al., Challenges for Nomadic Computing: Mobility Management and Wireless Communication, ACM/Baltzer Journal of Mobile Networking and Applications, Vol. 1, No. 1, 1996. M. Satyanarayan, Fundamental Challenges in Mobile Computing, M. Satyanarayan, Fifteen ACM Symposium on Principles of Distributed Computing, 1996. CS 515 İbrahim Körpeoğlu

Challenges of Mobile Computing and Networking
Challenges of Wireless/Mobile Network Design Challenges of Mobile System and Application Design We will look to the problems more from Computer Science point of view CS 515 İbrahim Körpeoğlu

Enabling Developments for Mobile Computing
Two factors that enabled mobile and ubiquitous computing (also called nomadic computing) Advances in wireless communication systems (both voice and data) and networks Flexible communication Less dependence on location for network access You are not limited with length of cable Advance in computer technology and development of portable computers and devices Wide use of laptop computers Introduction of Palmtop and hand-help computers CS 515 İbrahim Körpeoğlu

Factors challenging Mobile Computing
Wireless Communication Implications of using wireless communication for mobile computing The differences between wireless and wired media Mobility Consequences of mobility on mobile application and system design Portability Pressures that portability places in the design of mobile end-systems CS 515 İbrahim Körpeoğlu

Wireless network access is flexible and less location dependent Wireless communication is much more difficult to achieve than wired communication Wireless signals are affected by surrounding environment Blocking of the signals (walls etc.) Interference from other signal sources Reflections and fading CS 515 İbrahim Körpeoğlu

Wireless connections are of lower quality Lower bandwidths (bit-rates) Higher error-rates and burst errors More disconnections These factors increase the communication latency due to Losses and retransmissions Retransmission timeout delays Error control protocol processing Short disconnections CS 515 İbrahim Körpeoğlu

Wireless connections can be lost due to Mobility that results out of coverage area roaming Radio signal strengths drops with increasing distance between a wireless transmitter and receiver High interference at some locations Other devices around that use the same frequency band High load on some cells Lots of users who want to talk and access the network at the same time CS 515 İbrahim Körpeoğlu

Design Challenges Wireless Communication brings challenges to mobile computing because of Disconnections Low Bandwidth High Bandwidth Variability Heterogeneous Networks Security Risks Mobile Systems and Applications should consider these issues for good operation/functionality performance availability CS 515 İbrahim Körpeoğlu

Wireless Communication - Disconnections
Todays computers depend heavily on network Network File Systems, ftp servers, telnet serves, X-servers, Web servers Network failure will stall the applications and systems Network failure is greater concern for mobile computing Disconnections can be much more frequent CS 515 İbrahim Körpeoğlu

There is trade-off between autonomy and distributed computing The more autonomous the mobile computers, the better they can tolerate to network disconnections However, since mobile computer resources are scarce and limited, it is preferable to use the network and network services as much as possible to off-load computation and storage to network For example using a network file system prevents storing all the files in the local mobile computer CS 515 İbrahim Körpeoğlu

Code File System is a good example of handling network disconnections Designed as a file system for mobile computers like laptops Information from user profiles is used to locally cache best selection of files on the mobile computer A whole file is cached (not only some blocks) Optimistic caching scheme is used Users can update the cached copies Studies show that only rarely (1%) are files actually shared and written to in a distributed system When network reconnects, the cache is automatically reconciled with the master copy in the server. CS 515 İbrahim Körpeoğlu

Hoarding: Periodically a good set of files are copied from the master repository at the server to the mobile computer cache. The mobile users make their updates on the files All events are logged into a log file. When network reconnects, the log file is used to merge the updates and to make the caches consistent. CS 515 İbrahim Körpeoğlu

Wireless Communication – Low bandwidth
Mobile computing designs need to be more concerned about network bandwidth consumption and constraints than designs for stationary computing Wireless networks deliver lower bandwidth than wired networks 1 Mbps Infrared communication 11 Mbps wireless local radio communications (shared) 9.6 Kbps for wide-area wireless communication Wired networks Mbps for Ethernet 100 Mbps for FDDI 155 Mbps for ATM 1 Gbps for Gigabit Ethernet CS 515 İbrahim Körpeoğlu

To increase the system’s effective bandwidth per user 1) Use small cells with many base-stations OR 2) Use different frequencies with overlapping cells Weiser defined the capacity of wireless network as Bandwidth provided per cubic meter There is a hardware tradeoff between bandwidth and coverage area Transmitters covering a smaller area achieve higher bandwidths CS 515 İbrahim Körpeoğlu

Some software techniques to cope with low bandwidths Compress data that is to be transmitted Log the data, and use bulk transfers Bulk transfers are more efficient than many individual transfers in terms of bandwidth usage Lazy-write back of local caches of mobile computers may also reduce the network bandwidth demand Pre-fetching allows transferring the data ahead of need and thereby reduces the peek loads at time of many demands Scheduling packets on the wireless channels is also important. Priority should be given to packets that belong time-critical applications CS 515 İbrahim Körpeoğlu

Wireless Communication – High bandwidth variability
Mobile computers face much more variability in effective bandwidth than stationary computers Bandwidth can shift 1 to 4 orders of magnitude between wired and wireless communication A mobile application can cope with this bandwidth variability in 3 ways Assuming availability of high bandwidth connections and operating only on wired networks Assuming low bandwidth connections and not taking advantage of wired access and high bandwidths Adapting to the currently available bandwidth: providing the user with a variable level of quality and detail CS 515 İbrahim Körpeoğlu

Wireless Communication – Heterogeneous Networks
Stationary computers access the network over the same link for a long time No change in link characteristics: bandwidth, delay, loss-rate Mobile computers encounters heterogeneous network connections Using different base stations Some have better quality and less number of users Using different wireless technologies Indoor: infrared link; Outdoor: wide-area radio link Cities: cellular network; Rural areas: satellite network CS 515 İbrahim Körpeoğlu

Wireless Communication – Security Risks
It is much easier to connect to a wireless link than to connect to a wired link Two kind of security concerns Access control to wireless network You may not want other un-authorized people to access your wireless local area network at the company Use security protocols such as 802.1x that requires authentication of users to the Wireless LAN before they can transmit packets Prevent others to sniff and read the data packets that are sent over a wireless link Use encryption for data transmitted Shared keys are used (manual or automatic key management) CS 515 İbrahim Körpeoğlu

Mobility Mobility is ability to change locations while connected to the network This make the information more volatile A mobile computer may change the server that it is using when it moves to a new location The server could be for example a print server or a DNS server, etc. Main problems introduced by mobility Address migration Location Dependent Information Migrating locality CS 515 İbrahim Körpeoğlu

Mobility – Address Migration
In networks designed for static stations (Internet for example), an address has two functions It is used as the identity of the station It is also related with location of the station, hence is used for routing the packets to the station For supporting mobile hosts, the two functions need be separated We need a name for the mobile station that is independent of the current location of the mobile station We also need an address for the mobile station that shows the current location where it resides Adress_of_mobile = f (current_location_of_mobile) Hence the address changes when mobile station changes location CS 515 İbrahim Körpeoğlu

In order to communicate with a mobile computers, one needs to find it current location (address) Some methods to find the most recent address of a mobile computer are: Selective broadcast Central services Home bases Forwarding pointers CS 515 İbrahim Körpeoğlu

Selective Broadcast A message is sent to all cells in the networks Asking mobile computer to reply with its current address Too expensive for frequent use and queries Selectively directing the query to region or set of cells where the mobile is expected to be in Central Services The current address of each mobile computer is stored and maintained in a logically centralized database Each time mobile changes address, it updates the database with the new address The logically centralized database could be actually implemented using various common techniques: distribution, replication, and caching to improve both performance and availability CS 515 İbrahim Körpeoğlu

Home Bases A limited case of central service A single server at the home location of a mobile computer knows and maintains the current location of the mobile The location queries, or packets are first directed to this server at the home location Home location could be for example, the subnet indicated by the permanent IP address of a mobile computer Forwarding Pointers A pointer (the new current address) is kept at every location that mobile computers traverses. Chain of pointers could be too long No aging and removal (forgetting) mechanisms Requires an agent or entity at the old location to forward the packets to the new location CS 515 İbrahim Körpeoğlu

Mobility – Location Dependent Information
For classical stationary computers, the information that depends on the location is configured statically and usually manually The information include The IP addresses of the primary and secondary local DNS servers Available printers The time zone Mobile computers need to access also more location-dependent information Information about each room when you visit a museum Information about the current town/area when you travel with your car CS 515 İbrahim Körpeoğlu

Mobility – Location Dependent Information
Privacy Concerns The location information of a mobile user should not be revealed to everybody A burglar should not know the whereabouts of a home-owner The location information can be revealed in a controlled manner in some useful applications The location information of colleagues Routing telephone calls to the current location of a mobile user Tailoring the content of electronic bulletin boards depending on the mobile users that are roaming in the vicinity CS 515 İbrahim Körpeoğlu

Mobility – Migrating Locality
When the mobile moves, the distance between the mobile and services changes The physical distance is different then the network distance With a small change in physical location, network administrative domains could be crossed (from Bilkent Network to METU Network) Change in network distance may mean longer paths Longer latency Greater risk of disconnection More consumption of overall network capacity To avoid these, server connections may be dynamically transferred to servers that are closer CS 515 İbrahim Körpeoğlu

Portability Portability means that you can carry a computer or device
It is designed so that it is feasible and practical to carry it with you A mobile unit is not always portable A car is mobile but not portable (you can not carry it with you) A portable unit does not have to always mobile You can use your laptop always at home You can use your laptop at home or at school You can use your laptop at home, at school, and also while your are traveling at the campus ring bus or city bus. CS 515 İbrahim Körpeoğlu

Portability Desktop computers are not intended to be carried
Therefore their design is more liberal in their use of space, form-factor, power, cabling, and heat dissipation The design of a hand-held computer should strive for the following features: Small size Light-weight Durable (against dropping, hitting, etc) Water-resistant Long battery life-time Efficient in terms of screen use Have easy to use input devices CS 515 İbrahim Körpeoğlu

Portability Portability Constraints Include Low power consumption
You would not want to carry a battery that is bigger than your computer! Increased risk of data loss Small user-interfaces Limited on-board storage CS 515 İbrahim Körpeoğlu

Portability – Low power
Batteries are the largest source of weigth in a potable computer Minimizing power consumption can improve portability by reducign battery weight and lengthening the life of a charge Power consumption in a circuity P ~ CV2F P: power C: capacitance V: voltage (5V, 3V, etc.) F: clock speed CS 515 İbrahim Körpeoğlu

Power can be saved by design Reduce capacitance by greater levels of VLSI integration Reduce voltage by redesigning chips that operate at lower voltages Reduce clock frequency Trade-off between computational speed and power savings Design processor that do more work per clock cycle. CS 515 İbrahim Körpeoğlu

Power can be saved by operation Use power management software That powers down individual components when they are idle (disks, LCD screens for example) Applications can conserve power by reducing their appetite for computation, communication and memory Perform periodic operations more infrequently Trading talking for listening Transmission consumes more power than receiption in a communication device (10 times in cellular phones) CS 515 İbrahim Körpeoğlu

Power consumption breakdown by subsystems of a portable computer System % Power Display Edge-Light 35% CPU/Memory 31% Hard Disk 10% Floppy Disk 8% Display 5% Keyboard 1% CS 515 İbrahim Körpeoğlu

Portability – Risk to Data
Making computers portable increases their risk of Physical damage Unauthorized access Loss and Theft The risks can be reduced by minimizing essential data that is kept on board Make backup copies Prevent unauthorized disclosure of information Use encryption for the data that is stored on the disks and memory CS 515 İbrahim Körpeoğlu

Portability – Small User Interface
Size contstaints on portable computers require small user interface Requires a different windowing scheme (multpile windows are not appropriate) Head-mounted virtual reality displays Buttons versus Recognition There is not much space for a full keyboard Trade buttons in favor of recognizing user’s intentions from analog input devices Handwriting recognition (96-98% accuracy) Voice recognition (96-98% accuracy) Storage and processing demand Disturbs others, compromise privacy Gesture recognition CS 515 İbrahim Körpeoğlu

Portability – Small User Interface
Pointing Devices Mouse does not suite for mobile computers Switch to Pens Requires change in user interface and also in software interface Pens can jump to any location Pens can be used for writing besides pointing Pen positioning resolution is several times that of screen resolution CS 515 İbrahim Körpeoğlu

Small Storage Capacity
Storage capacity is limited because of physical size and power requirements Portable devices do not use disks They consume too much power They can not endure to the un-nice treartment that most portable computers face Coping with limited storage Copressing file systems Accessing remote storage over the network Sharing code libraries Compressing virtual memory pages Using interpreted languages instead of translated (compiled) languages We don’t need object code in this case Object code is many times larger than the source code CS 515 İbrahim Körpeoğlu

Summary Wireless communication brings challenging network conditions Slow and sometime disconnected communication Mobility causes greater dynamicisim of information Portability results limited resources to be available on board Mobile computing designers should consider these issues in designing mobile systems, applications and networks that are comparable with the traditional stationary computing and communication in terms of operation, performance, and availability CS 515 İbrahim Körpeoğlu

General Techniques to face Challenges

Techniques to face Challenges
Three general techniques that have been applied in various systems for supporting mobile/nomadic computing Asymmetric design of protocols and applications To overcome the limitations of mobile devcices Use of Network based proxies To perform computing and communication functions on behalf of mobile users Use of intelligent caching and prefetching techniques To imrove performance and availability CS 515 İbrahim Körpeoğlu

Outline First discuss these general techniques
Network bases proxies Judicious acquisition and caching of Information Asymmetric Protocols and Applications The solutions are not limited to these techniques Then we will describe example systems that make use of these techniques CS 515 İbrahim Körpeoğlu

General Solution Techniques – Network Based Proxies
Many mobile systems make use of intelligent agents that reside inside the wired network and perform various functions on behalf of mobile users Cellular systems use intelligent switches and databases that store user profiles to perform functions on behalf of these users Intelligent agents, called also proxies, can be used to process control information (take part in connection establishment/termination for the user) Proxies are also used to manipulate user information that is being exchanged between the mobile device and a network-based server. CS 515 İbrahim Körpeoğlu

General Solution Techniques – Network Based Proxies
General Benefits of Network-based Proxies Proxies may execute complex functions relieving processing limited mobile devices. Proxies may be used the reduce the amount of communication required with the mobile device thus reducing the amount of air interface bandwidth consumed. Proxies may account for mobile devices that are in disconnected state. Proxies may shield network-based applications from the mobility of their clients. Proxies may shield applications from the heterogeneity of mobile devices. CS 515 İbrahim Körpeoğlu

Example Functions of Proxies
Format Translation The information sent from servers to mobile devices is reformatted A postscript file could be converted to an ascii text file at the proxy if the mobile device display can only support text In this way: The translation overhead (processing + storage) is relieved from mobile device to the server The bandwidth demand on the wireless link between mobile device and proxy is reduced. The web server need not to be modified to support different kind of mobile devices (the proxy handles them) CS 515 İbrahim Körpeoğlu

Control Functions Circuit-oriented communication requires connection establishment (signaling) before data/voice is transmitted. Signaling functions: Negotiation application and network capabilities Allocating resources Signaling involves a lot of message exchanges Proxies may help in signaling and reduce the signaling messages that the mobile device has to send and receive CS 515 İbrahim Körpeoğlu

Filter/Modify Application Information The images in a web page may be filtered out at the proxy if the mobile device can only support text output The frame quality of a video stream can be reduced so that it is transferred with a lower bandwidth demand over the wireless link The video streams can be trans-coded into different compression schemes in order to adapt the bandwidth the video requires to the available bandwidth on the wireless communication channel. CS 515 İbrahim Körpeoğlu

Account for mobile devices Proxies can also account for disconnected or powered-off mobile devices The SMS messages that are sent to cellular phone subscriber can be stored in the network for later retrievel if the subscriber can not be reached (either because of powering-off the mobile unit or because of out-of-range traveling). This overcomes problems associated with limited coverage to support reliable communication. CS 515 İbrahim Körpeoğlu

Hiding Mobility The proxies can be used to hide the mobility of the users from the correspondents (other users or servers). Proxy knows how to reach the mobile user Disadvantage is that all information should be intervened and processed at the proxy This could degrade the performance CS 515 İbrahim Körpeoğlu

Hiding Heterogeneity Proxies perform conversion functions depending on the capabilities of the mobile devices, the standards they conform, etc. This way the correspondent application need not to be aware of the different characteristics of the mobile end-devices CS 515 İbrahim Körpeoğlu

General Solution Techniques – Pre-fetching and Caching of Information
These techniques are used in mobile computing for: Limiting the communication caused by mobility Improving performance and availability of services CS 515 İbrahim Körpeoğlu

Example of Pre-fetching and Caching
Location information The information about the current location of a mobile device could be cached at a server This limits the amount of control traffic required in the network to locate a mobile device Reduce signaling overhead Accurate caching of location information reduces also the time taken to locate a mobile device Reduce delay Frequency of location updates If too frequent, control messages will occupy the network and will be significant portion of the total traffic If not frequent enough, the data cached could be stale and performance may degrade CS 515 İbrahim Körpeoğlu

Example of Pre-fetching and Caching
The mobile unit can pre-fetch items (files, etc) at low-load network conditions The pre-fetching is done in the background for non-real time applications Prevent the network overload when a lot of requests on the data items are close in the time. CS 515 İbrahim Körpeoğlu

General Solution Techniques – Asymmetric Protocols and Applications
Asymmetric protocol and application design helps to overcome the inherent imbalances of: Processing power between the mobile wireless end devices and network-based processors Uplink and downlink bandwidth available due to transmission power available from wireless mobile devices CS 515 İbrahim Körpeoğlu

Asymmetric Protocols and Applications
Lower layer protocols can be developed that place higher processing and memory requirements on fixed servers New applications can be developed for mobile environment The applications will take into account the asymmetric nature of wireless communication links and also processing elements. CS 515 İbrahim Körpeoğlu

General Solution Techniques - System Examples

Application of Techniques
On Error Control at the Data-Link Layer On Routing at the Network Layer On Some Applications CS 515 İbrahim Körpeoğlu

Error Control Wireless link experience much higher error rates compared to wired links, because of Wireless Channels characteristics such as fading, attenuation, interference, etc. Mobility A user moving out of range of a base station or moving into a crowded area Data can be lost during handoff Errors incurred on the wireless link have dramatic effects on the performance of reliable transport protocols. CS 515 İbrahim Körpeoğlu

Error Control Effect of errors on TCP
TCP assumes packet losses are due to congestion which usually takes a long period of time Hence it triggers congestion recovery procedures when packet losses occurs This slows down TCP dramatically and takes quite a lot time to speedup again However, on a wireless link packet losses are due link errors which take a short period of time There is a need to know if the packet losses are due to congestion or due to link error to react appropriately CS 515 İbrahim Körpeoğlu

Error Control Two approaches to recover from wireless link errors
Use reliable link layer protocols More complex link layer May interact with TCP reliability, timers, etc. Let the TCP recover from packet losses to due link error Simpler link layer TCP source may need to distinguish packet losses due to congestion and due to link layer errors CS 515 İbrahim Körpeoğlu

Error Control – Link layer solutions
There are methods to alleviate the packet losses on a wireless link due to high error rate Use of FEC (Forward Error Correction) Schemes Increases the packet length with redundant bits to correct the bit errors on the link No retransmissions Use of ARQ (Automatic Repeat Request) Schemes Does not affect the original packet length Retransmits the packets that are lost Need feedback from the receiver if the packet(s) received successfully or not Need retransmission timers to detect losses if the receiver does not send negative acknowledgements CS 515 İbrahim Körpeoğlu

Error Control – Use of Asymmetry
AIRMAIL protocol designed at Lucent/Bell Labs, Holmdel, New Jersey. A link layer protocol designed to work over wireless links Two key ideas: Provides reliable delivery over wireless link so that the TCP like protocols are not affected by high link error rates Support mobility so that reliable delivery is ensured without a large performance degradation during handoffs. Provides reliable delivery using a combination of ARQ and FEC techniques. FEC technique adapts to the raw bit error rate that is experienced over the wireless link CS 515 İbrahim Körpeoğlu

Error Control – ARQ Basics of ARQ Mobile Device Base Station
Packet N Ack N+1 Sequence Numbers are used for: In order delivery of packets To eliminate duplicate packets Packet N+1 lost Retransmission Timeout Packet N+1 CS 515 İbrahim Körpeoğlu

Error Control – Asymmetric ARQ
AIRMAIL uses Asymmetric ARQ Place majority of processing complexity at the protocol entity based inside the wired network at the base-station Place no-timers on the mobile device Reduce the processing required due to acknowledgments at the mobile device CS 515 İbrahim Körpeoğlu

Error Control – AIRMAIL Asymmetric ARQ
Communication from base station to mobile Base station maintains retransmission timers Mobile device generates block acknowledgements unless specifically requested by the base station No timer required at the mobile Number of ACKs generated is limited Communication from mobile to base station When mobile transmits packets, it maintains a table where it stores the time at which each packet is transmitted Base station periodically transmits the status of its receiver to the mobile. The mobile checks the receive time of status message against the transmission times stored in the packet records and determines if the packet has not received an ACK in more than one round-trip-time (RTT). If so, these packets are retransmitted. This way, mobile has no timers and it receives ACKs periodically. CS 515 İbrahim Körpeoğlu

Error Control – AIRMAIL
AIRMAIL advantages Compiled software size Base station: 150 Kbytes Mobile unit: 100 Kbytes 2/3 code reduction Processing time to transmit 200 Kbytes of data Base station: 0.7 seconds Mobile unit: 0.23 seconds 1/3 reduction in processing time CS 515 İbrahim Körpeoğlu

Error Control: Use of Proxies
Proxies can also be used at base station to overcome errors on the wireless link SNOOP protocol designed at UC, Berkeley is an example that improves the performance of TCP over wireless links SNOOP acts as a proxy that is located on a base station over the path of the TCP connection between a mobile device and a correspondent host (fixed host). CS 515 İbrahim Körpeoğlu

SNOOP SNOOP monitors every TCP segment that is sent to and received from a mobile host. It shields the TCP at the fixed host from experiencing the effect of data loss on the wireless link CS 515 İbrahim Körpeoğlu

SNOOP TCP TCP Mobile Host Fixed Host Base Station SNOOP CS 515
İbrahim Körpeoğlu

SNOOP Operation From Fixed Host to Mobile
SNOOP intercepts and monitors at the Base Station every TCP segment that is sent from Fixed Host to the Mobile Host Each segment is stored in a Cache. SNOOP also monitors all the ACKs that are sent from the Mobile Host to the Fixed Host When a normal ACK is received at SNOOP, it removes all segments up-to that ACK number and sends an ACK back to the fixed host When a duplicate ACK is received at SNOOP (indication of packet loss), SNOOP retransmit the segment from its cache. The duplicate ACK is not sent back to the fixed host Fixed host does not trigger congestion control procedures CS 515 İbrahim Körpeoğlu

SNOOP Operation From Mobile Host to Fixed Host
SNOOP also monitors all TCP segments that are sent from mobile host to the fixed host. If it detects that segments are lost on the wireless link, it generates negative acknowledgments to the mobile host and recovery starts at the mobile host. For BER of 2x10-6 on wireless channel SNOOP provides around 67% improvement in throughput over not using SNOOP. CS 515 İbrahim Körpeoğlu

Routing In static Internet and telephone network, the end-device address designates its location and also its identity. For mobile network, this is no longer valid. Identity and address needs to be seperated Identity remains fixed wherever the mobile moves Address changes depending on the current location of the mobile One problem is how to locate the mobile and how to route the packets (or circuits) to it. Caching of location information and use of proxies can help with this. CS 515 İbrahim Körpeoğlu

Routing and Caching Caching of location information is used in varying degrees in telecommunication, PCS, and packet networks Cellular Telecom Network use of caches A connection is established before voice/data is transmitted A mobile device need to be located in order to route the connection request to it and then establish the connection Cellular Networks use two-tier location database structure for location discovery procedures CS 515 İbrahim Körpeoğlu

GSM Location Tracking and Call Setup
All subscriber info is kept at HLR HLR keeps the mobile-VLR binding HLR 5 – De-registration 3 - Registration VLR keeps track of the mobiles in its Region (mobile-address Binding) 2 – Learn about HLR Old VLR New VLR Two-tier hierarchy of Databases: HLR: Home Location Register VLR: Visitor Location Register 1 4 Mobile Host CS 515 İbrahim Körpeoğlu

GMSC 1 1 HLR VLR 2 2 1 1 3 MSC 3 Switches Switches The identity of the VLR that is serving the mobile can be cached at the home switch (GMSC); In this way, HLR can be by-passed. CS 515 İbrahim Körpeoğlu

Location information of mobile devices is stored in VLRs. A VLR serves a well-defined region. Although it does have to be always true, we can say that generally there is one VLR per MSC serving a well-defined area. HLR keeps all the subscriber information and also the current VLR identity that is serving the mobile device HLR queries the VLR to find out the routable address of the mobile device. CS 515 İbrahim Körpeoğlu

Routing in Packet Networks – Mobile IP
The location information of a mobile device is cached at a home server, called home agent. When a mobile moves into a foreign location, it registers with a foreign agent. The foreign agent then registers the mobile device with the home agent Thereby, home agent knows which foreign agent serves the mobile device at current time. CS 515 İbrahim Körpeoğlu

Mobile IP – Routing Paths
Correspondent Host (A host that wants to communicate with the Mobile Host) Foreign Agent Internet Home Agent (Knows the current Location of the Mobile Host) Mobile Host Route Optimization can be achieved by caching the current location of the mobile device at the correspondent host CS 515 İbrahim Körpeoğlu

Applications Use of Solution Techniques on the following areas:
File Systems Multimedia Networking Web Browsing CS 515 İbrahim Körpeoğlu

Caching: File Systems Coda is an example file system, which uses caching for: Supporting disconnected mode of operation over a network file system designed for nomadic computers Increases system availability Increasing the performance of the file system We described Coda briefly on earlier slides A web browsing application (W4) uses pre-fetching and caching A mobile PDA is mated with a proxy in the network Proxy caches the pages and sends them to PDA upon request Subsequent pages are pre-fetched and cached at the PDA Cached access from PDA is around 1sec, proxy access is around 2-3sec. CS 515 İbrahim Körpeoğlu

Summary Characteristics of Nomadic Computing environment is described
How does it affect the mobile system and application design General Solution Techniques to face the challenges of mobile computing is described Some example systems that use these techniques are introduced. CS 515 İbrahim Körpeoğlu

Radio Propagation - Large-Scale Path Loss
CS 515 Mobile and Wireless Networking Ibrahim Korpeoglu Computer Engineering Department Bilkent University, Ankara

Reading Homework Due Date: Next Wednesday
Read the handouts that wa have put into the library Read the following paper: J. B. Andersen, T. S. Rappaport, S. Yoshida, Propagation Measurements and Models for Wireless Communications Channels, IEEE Communications Magazine, (January 1995), pp Read Chapter 4 of Rappaport’s Book. You have of course read the previous 4 papers! CS 515 Ibrahim Korpeoglu

Basics When electrons move, they create electromagnetic waves that can propagate through the space Number of oscillations per second of an electromagnetic wave is called its frequency, f, measured in Hertz. The distance between two consecutive maxima is called the wavelength, designated by l. CS 515 Ibrahim Korpeoglu

Basics By attaching an antenna of the appropriate size to an electrical circuit, the electromagnetic waves can be broadcast efficiently and received by a receiver some distance away. In vacuum, all electromagnetic waves travel at the speed of light: c = 3x108 m/sec. In copper or fiber the speed slows down to about 2/3 of this value. Relation between f, l , c: lf = c CS 515 Ibrahim Korpeoglu

Basics We have seen earlier the electromagnetic spectrum.
The radio, microwave, infrared, and visible light portions of the spectrum can all be used to transmit information By modulating the amplitude, frequency, or phase of the waves. CS 515 Ibrahim Korpeoglu

Basics We have seen wireless channel concept earlier: it is characterized by a frequency band (called its bandwidth) The amount of information a wireless channel can carry is related to its bandwidth Most wireless transmission use narrow frequency band (Df << f) Df: frequency band f: middle frequency where transmission occurs New technologies use spread spectrum techniques A wider frequency band is used for transmission CS 515 Ibrahim Korpeoglu

Basics - Propagation Radio waves are Easy to generate
Can travel long distances Can penetrate buildings They are both used for indoor and outdoor communication They are omni-directional: can travel in all directions They can be narrowly focused at high frequencies (greater than 100MHz) using parabolic antennas (like satellite dishes) Properties of radio waves are frequency dependent At low frequencies, they pass through obstacles well, but the power falls off sharply with distance from source At high frequencies, they tend to travel in straight lines and bounce of obstacles (they can also be absorbed by rain) They are subject to interference from other radio wave sources CS 515 Ibrahim Korpeoglu

Basics - Propagation At VLF, LF, and MF bands, radio
waves follow the ground. AM radio broadcasting uses MF band reflection At HF bands, the ground waves tend to be absorbed by the earth. The waves that reach ionosphere ( km above earth surface), are refracted and sent back to earth. Ionosphere absorption CS 515 Ibrahim Korpeoglu

Basics - Propagation VHF Transmission LOS path Reflected Wave
Directional antennas are used Waves follow more direct paths - LOS: Line-of-Sight Communication - Reflected wave interfere with the original signal CS 515 Ibrahim Korpeoglu

Basics - Propagation Waves behave more like light at higher frequencies Difficulty in passing obstacles More direct paths They behave more like radio at lower frequencies Can pass obstacles CS 515 Ibrahim Korpeoglu

Propagation Models We are interested in propagation characteristics and models for waves with frequencyy in range: few MHz to a few GHz Modeling radio channel is important for: Determining the coverage area of a transmitter Determine the transmitter power requirement Determine the battery lifetime Finding modulation and coding schemes to improve the channel quality Determine the maximum channel capacity CS 515 Ibrahim Korpeoglu

Radio Propagation Models
Transmission path between sender and receiver could be Line-of-Sight (LOS) Obstructed by buildings, mountains and foliage Even speed of motion effects the fading characteristics of the channel CS 515 Ibrahim Korpeoglu

Radio Propagation Mechanisms
The physical mechanisms that govern radio propagation are complex and diverse, but generally attributed to the following three factors Reflection Diffraction Scattering Occurs when waves impinges upon an obstruction that is much larger in size compared to the wavelength of the signal Example: reflections from earth and buildings These reflections may interfere with the original signal constructively or destructively CS 515 Ibrahim Korpeoglu

Diffraction Occurs when the radio path between sender and receiver is obstructed by an impenetrable body and by a surface with sharp irregularities (edges) Explains how radio signals can travel urban and rural environments without a line-of-sight path Scattering Occurs when the radio channel contains objects whose sizes are on the order of the wavelength or less of the propagating wave and also when the number of obstacles are quite large. They are produced by small objects, rough surfaces and other irregularities on the channel Follows same principles with diffraction Causes the transmitter energy to be radiated in many directions Lamp posts and street signs may cause scattering CS 515 Ibrahim Korpeoglu

transmitter Street S D D Building Blocks R: Reflection D: Diffraction S: Scattering receiver CS 515 Ibrahim Korpeoglu

As a mobile moves through a coverage area, these 3 mechanisms have an impact on the instantaneous received signal strength. If a mobile does have a clear line of sight path to the base-station, than diffraction and scattering will not dominate the propagation. If a mobile is at a street level without LOS, then diffraction and scattering will probably dominate the propagation. CS 515 Ibrahim Korpeoglu

As the mobile moves over small distances, the instantaneous received signal will fluctuate rapidly giving rise to small-scale fading The reason is that the signal is the sum of many contributors coming from different directions and since the phases of these signals are random, the sum behave like a noise (Rayleigh fading). In small scale fading, the received signal power may change as much as 3 or 4 orders of magnitude (30dB or 40dB), when the receiver is only moved a fraction of the wavelength. CS 515 Ibrahim Korpeoglu

As the mobile moves away from the transmitter over larger distances, the local average received signal will gradually decrease. This is called large-scale path loss. Typically the local average received power is computed by averaging signal measurements over a measurement track of 5l to 40l. (For PCS, this means 1m-10m track) The models that predict the mean signal strength for an arbitrary-receiver transmitter (T-R) separation distance are called large-scale propagation models Useful for estimating the coverage area of transmitters CS 515 Ibrahim Korpeoglu

Small-Scale and Large-Scale Fading
Received Power (dBm) -30 -40 -50 -60 This figure is just an illustration to show the concept. It is not based on read data. -70 T-R Separation (meters) CS 515 Ibrahim Korpeoglu

What is Decibel (dB) What is dB (decibel):
A logarithmic unit that is used to describe a ratio. Let say we have two values P1 and P2. The difference (ratio) between them can be expressed in dB and is computed as follows: 10 log (P1/P2) dB Example: transmit power P1 = 100W, received power P2 = 1 W The difference is 10log(100/1) = 20dB. CS 515 Ibrahim Korpeoglu

dB dB unit can describe very big ratios with numbers of modest size.
See some examples: Tx power = 100W, Received power = 1W Tx power is 100 times of received power Difference is 20dB Tx power = 100W, Received power = 1mW Tx power is 100,000 times of received power Difference is 50dB Tx power = 1000W, Received power = 1mW Tx power is million times of received power Difference is 60dB CS 515 Ibrahim Korpeoglu

dBm For power differences, dBm is used to denote a power level with respect to 1mW as the reference power level. Let say Tx power of a system is 100W. Question: What is the Tx power in unit of dBm? Answer: Tx_power(dBm) = 10log(100W/1mW) = 10log(100W/0.001W) = 10log(100,0000) = 50dBm CS 515 Ibrahim Korpeoglu

dBW For power differences, dBW is used to denote a power level with respect to 1W as the reference power level. Let say Tx power of a system is 100W. Question: What is the Tx power in unit of dBW? Answer: Tx_power(dBW) = 10log(100W/1W) = 10log(100) = 20dBW. CS 515 Ibrahim Korpeoglu

Decibel (dB) versus Power Ratio Comparison of two Sound Systems CS 515
Ibrahim Korpeoglu

Free-Space Propagation Model
Used to predict the received signal strength when transmitter and receiver have clear, unobstructed LOS path between them. The received power decays as a function of T-R separation distance raised to some power. Path Loss: Signal attenuation as a positive quantity measured in dB and defined as the difference (in dB) between the effective transmitter power and received power. CS 515 Ibrahim Korpeoglu

Free space power received by a receiver antenna separated from a radiating transmitter antenna by a distance d is given by Friis free space equation: Pr(d) = (PtGtGrl2) / ((4p)2d2L) [Equation 1] Pt is transmited power Pr(d) is the received power Gt is the trasmitter antenna gain (dimensionless quantity) Gr is the receiver antenna gain (dimensionless quantity) d is T-R separation distance in meters L is system loss factor not related to propagation (L >= 1) L = 1 indicates no loss in system hardware (for our purposes we will take L = 1, so we will igonore it in our calculations). l is wavelength in meters. CS 515 Ibrahim Korpeoglu

The gain of an antenna G is related to its affective aperture Ae by: G = 4pAe / l [Equation 2] The effective aperture of Ae is related to the physical size of the antenna, l is related to the carrier frequency by: l = c/f = 2pc / wc [Equation 3] f is carrier frequency in Hertz wc is carrier frequency in radians per second. c is speed of light in meters/sec CS 515 Ibrahim Korpeoglu

An isotropic radiator is an ideal antenna that radiates power with unit gain uniformly in all directions. It is as the reference antenna in wireless systems. The effective isotropic radiated power (EIRP) is defined as: EIRP = PtGt [Equation 4] Antenna gains are given in units of dBi (dB gain with respect to an isotropic antenna) or units of dBd (dB gain with respect to a half-wave dipole antenna). Unity gain means: G is 1 or 0dBi CS 515 Ibrahim Korpeoglu

Path loss, which represents signal attenuation as positive quantity measured in dB, is defined as the difference (in dB) between the effective transmitted power and the received power. PL(dB) = 10 log (Pt/Pr) = -10log[(GtGrl2)/(4p)2d2] [Equation 5] (You can drive this from equation 1) If antennas have unity gains (exclude them) PL(dB) = 10 log (Pt/Pr) = -10log[l2/(4p)2d2] [Equation 6] CS 515 Ibrahim Korpeoglu

For Friis equation to hold, distance d should be in the far-field of the transmitting antenna. The far-field, or Fraunhofer region, of a transmitting antenna is defined as the region beyond the far-field distance df given by: df = 2D2/l [Equation 7] D is the largest physical dimension of the antenna. Additionally, df >> D and df >> l CS 515 Ibrahim Korpeoglu

Free-Space Propagation Model – Reference Distance d0
It is clear the Equation 1 does not hold for d = 0. For this reason, models use a close-in distance d0 as the receiver power reference point. d0 should be >= df d0 should be smaller than any practical distance a mobile system uses Received power Pr(d), at a distance d > d0 from a transmitter, is related to Pr at d0, which is expressed as Pr(d0). The power received in free space at a distance greater than d0 is given by: Pr(d) = Pr(d0)(d0/d)2 d >= d0 >= df [Equation 8] CS 515 Ibrahim Korpeoglu

Expressing the received power in dBm and dBW Pr(d) (dBm) = 10 log [Pr(d0)/0.001W] + 20log(d0/d) where d >= d0 >= df and Pr(d0) is in units of watts [Equation 9] Pr(d) (dBW) = 10 log [Pr(d0)/1W] + 20log(d0/d) where d >= d0 >= df and Pr(d0) is in units of watts [Equation 10] Reference distance d0 for practical systems: For frequncies in the range 1-2 GHz 1 m in indoor environments 100m-1km in outdoor environments CS 515 Ibrahim Korpeoglu

Example Question A transmitter produces 50W of power.
A) Express the transmit power in dBm B) Express the transmit power in dBW C) If d0 is 100m and the received power at that distance is mW, then find the received power level at a distance of 10km. Assume that the transmit and receive antennas have unity gains. CS 515 Ibrahim Korpeoglu

Solution A) B) Pt(W) is 50W.
Pt(dBm) = 10log[Pt(mW)/1mW)] Pt(dBm) = 10log(50x1000) Pt(dBm) = 47 dBm B) Pt(dBW) = 10log[Pt(W)/1W)] Pt(dBW) = 10log(50) Pt(dBW) = 17 dBW CS 515 Ibrahim Korpeoglu

Solution Pr(d) = Pr(d0)(d0/d)2
Substitute the values into the equation: Pr(10km) = Pr(100m)(100m/10km)2 Pr(10km) = mW(10-4) Pr(10km) = 3.5x10-10W Pr(10km) [dBm] = 10log(3.5x10-10W/1mW) = 10log(3.5x10-7) = dBm CS 515 Ibrahim Korpeoglu

Two main channel design issues
Communication engineers are generally concerned with two main radio channel issues: Link Budged Design Link budget design determines fundamental quantities such as transmit power requirements, coverage areas, and battery life It is determined by the amount of received power that may be expected at a particular distance or location from a transmitter Time dispersion It arises because of multi-path propagation where replicas of the transmitted signal reach the receiver with different propagation delays due to the propagation mechanisms that are described earlier. Time dispersion nature of the channel determines the maximum data rate that may be transmitted without using equalization. CS 515 Ibrahim Korpeoglu

Link Budged Design Using Path Loss Models
Radio propagation models can be derived By use of empirical methods: collect measurement, fit curves. By use of analytical methods Model the propagation mechanisms mathematically and derive equations for path loss Long distance path loss model Empirical and analytical models show that received signal power decreases logarithmically with distance for both indoor and outdoor channels CS 515 Ibrahim Korpeoglu

Announcements Please download the homework from the course webpage again. I made some important corrrections and modifications! I recommend that you read Chapter 4 of the book and “radio propagation paper” in parallel. Start early doing the homework! The last night before deadline may be loo late!!!. I put some links about math and statistics resources on the course webpage. I put a Z table on the webpage. Q-table and Z-table are related as follows: Q_table(z) = Z_table(z) CS 515 Ibrahim Korpeoglu

Long distance path loss model
The average large-scale path loss for an arbitrary T-R separation is expressed as a function of distance by using a path loss exponent n: The value of n depends on the propagation environment: for free space it is 2; when obstructions are present it has a larger value. Equation 11 CS 515 Ibrahim Korpeoglu

Path Loss Exponent for Different Environments
Path Loss Exponent, n Free space 2 Urban area cellular radio 2.7 to 3.5 Shadowed urban cellular radio 3 to 5 In building line-of-sight 1.6 to 1.8 Obstructed in building 4 to 6 Obstructed in factories 2 to 3 CS 515 Ibrahim Korpeoglu

Selection of free space reference distance
In large coverage cellular systems 1km reference distances are commonly used In microcellular systems Much smaller distances are used: such as 100m or 1m. The reference distance should always be in the far-field of the antenna so that near-field effects do not alter the reference path loss. CS 515 Ibrahim Korpeoglu

Log-normal Shadowing Equation 11 does not consider the fact the surrounding environment may be vastly different at two locations having the same T-R separation This leads to measurements that are different than the predicted values obtained using the above equation. Measurements show that for any value d, the path loss PL(d) in dBm at a particular location is random and distributed normally. CS 515 Ibrahim Korpeoglu

Log-normal Shadowing- Path Loss
Then adding this random factor: Equation 12 denotes the average large-scale path loss (in dB) at a distance d. Xs is a zero-mean Gaussian (normal) distributed random variable (in dB) with standard deviation s (also in dB). is usually computed assuming free space propagation model between transmitter and d0 (or by measurement). Equation 12 takes into account the shadowing affects due to cluttering on the propagation path. It is used as the propagation model for log-normal shadowing environments. CS 515 Ibrahim Korpeoglu

Log-normal Shadowing- Received Power
The received power in log-normal shadowing environment is given by the following formula (derivable from Equation 12) The antenna gains are included in PL(d). Equation 12 CS 515 Ibrahim Korpeoglu

Log-normal Shadowing, n and s
The log-normal shadowing model indicates the received power at a distance d is normally distributed with a distance dependent mean and with a standard deviation of s In practice the values of n and s are computed from measured data using linear regression so that the difference between the measured data and estimated path losses are minimized in a mean square error sense. CS 515 Ibrahim Korpeoglu

Example of determining n and s
Assume Pr(d0) = 0dBm and d0 is 100m Assume the receiver power Pr is measured at distances 100m, 500m, 1000m, and 3000m, The table gives the measured values of received power Distance from Transmitter Received Power 100m 0dBm 500m -5dBm 1000m -11dBm 3000m -16dBm CS 515 Ibrahim Korpeoglu

We know the measured values. Lets compute the estimates for received power at different distances using long-distance path loss model. (Equation 11) Pr(d0) is given as 0dBm and measured value is also the same. mean_Pr(d) = Pr(d0) – mean_PL(from_d0_to_d) Then mean_Pr(d) = 0 – 10logn(d/d0) Use this equation to computer power levels at 500m, 1000m, and 3000m. CS 515 Ibrahim Korpeoglu

Average_Pr(500m) = 0 – 10logn(500/100) = -6.99n Average_Pr(1000m) = 0 – 10logn(1000/100) = -10n Average_Pr(3000m) = 0 – 10logn(3000/100) = n Now we know the estimates and also measured actual values of the received power at different distances In order approximate n, we have to choose a value for n such that the mean square error over the collected statistics is minimized. CS 515 Ibrahim Korpeoglu

Example of determining n and s: MSE(Mean Square Error)
The mean square error (MSE) is given with the following formula: [Equation 14] Since power estimate at some distance depends on n, MSE(n) is a function of n. We would like to find a value of n that will minimize this MSE(n) value. We We will call it MMSE: minimum mean square error. This can be achieved by writing MSE as a function of n. Then finding the value of n which minimizes this function. This can be done by derivating MSE(n) with respect to n and solving for n which makes the derivative equal to zero. CS 515 Ibrahim Korpeoglu

Example of determining n:
Distance Measured Value of Pr (dBm) Estimated Value of Pr (dBm) 100m 500m -5 -6.99n 1000m -11 -10n 3000m -16 -14.77n MSE = (0-0)2 + (-5-(-6.99n))2 + (-11-(-10n)2 + (-16-(-14.77n)2 MSE = 0 + (6.99n – 5)2 + (10n – 11)2 + (14.77n – 16)2 If we open this, we get MSE as a function of n which as second order polynomial. We can easily take its derivate and find the value of n which minimizes MSE. ( I will not show these steps, since they are trivial). CS 515 Ibrahim Korpeoglu

Example of determining s:
We are interested in finding the standard deviation about the mean value For this, we will use the following formula Equation 14.1 Equation 14.2 CS 515 Ibrahim Korpeoglu

Some Statistics Knowledge: Computation of mean (m), variance (s2) and standard deviation (s)
Assume we have k samples (k values) X1, X2, …, Xk : The mean is denoted by m. The variance is denotes by s. The standard deviation is denotes by s2. The formulas to computer m, s, and s2 is given below: [Equation 15] [Equation 16] [Equation 17] CS 515 Ibrahim Korpeoglu

Path loss and Received Power
In log normal shadowing environment: PL(d) (path loss) and Pr(d) (received power at a distance d) are random variables with a normal distribution in dB about a distance dependent mean. Sometime we are interested in answering following kind of questions: What is mean received Pr(d) power (mean_Pr(d))at a distance d from a transmitter What is the probability that the receiver power Pr(d) (expressed in dB power units) at distance d is above (or below) some fixed value g (again expressed in dB power units such as dBm or dBW). CS 515 Ibrahim Korpeoglu

Received Power and Normal Distribution
In answering these kind of question, we have to use the properties of normal (gaussian distribution). Pr(d) is normally distributed that is characterized by: a mean (m) a standard deviation (s) We are interested in Probability that Pr(d) >= g or Pr(d) <= g CS 515 Ibrahim Korpeoglu

Received Power and Normal Distribution PDF
Figure shows the PDF of a normal distribution for the received power Pr at some fixed distance d ( m = 10, s = 5) (x-axis is received power, y-axis probability) EXAMPLE: Probability that Pr is smaller than 3.3 (Prob(Pr <= 3.3)) is given with value of the stripped area under the curve. CS 515 Ibrahim Korpeoglu

Normal CDF The figure shows the CDF plot of the normal distribution described previously. Prob(Pr <= 3.3) can be found by finding first the point where vertical line from 3.3 intersects the curve and then by finding the corresponding point on the y-axis. This corresponds to a value of Hence Prob(Pr <= 3.3) = 0.09 0.5 CS 515 Ibrahim Korpeoglu

Use of Normal Distribution
[Equation 18] PDF (probability density function of a normal distribution is characterized by two parameters, m (mean)and s (standard deviation), and given with the formula above. CS 515 Ibrahim Korpeoglu

To find out the probability that a Gaussian (normal) random variable X is above a value x0, we have to integrate pdf. Equation 19 This integration does not have any closed form. Any Gaussian PDF can be rewritten through substitution of y = x–m / s to yield Equation 20 CS 515 Ibrahim Korpeoglu

In the above formula, the kernel of the integral is normalized Gaussian PDF function with m = 0 and s = 1. Evaluation of this function is designed as Q-function and defined as Equation 21 Hence Equation 19 or 20 can be evaluated as: Equation 22 CS 515 Ibrahim Korpeoglu

Q-Function Q-Function is bounded by two analytical expressions as follows: Equation 23 For values greater than 3.0, both of these bounds closely approximate Q(z). Two important properties of Q(z) are: Q(-z) = 1 – Q(z) Equation 24 Q(0) = 1/2 Equation 25 CS 515 Ibrahim Korpeoglu

Tabulation of Q-function (0<=z<=3.9)
Q(z) 0.0 0.5 1.0 2.0 3.0 0.1 1.1 2.1 3.1 0.2 1.2 2.2 3.2 0.3 1.3 2.3 3.3 0.4 1.4 2.4 3.4 1.5 2.5 3.5 0.6 1.6 2.6 3.6 0.7 1.7 2.7 3.7 0.8 1.8 2.8 3.8 0.9 1.9 2.9 3.9 For values of z higher than 3.9, you should use the equations on the previous slide to compute Q(z). CS 515 Ibrahim Korpeoglu

Q-Function Graph: z versus Q(z)
z (1 <= z <= 3.9 CS 515 Ibrahim Korpeoglu

Erf and Erfc functions The error function (erf) is defined as:
[Equation 26] And the complementary error function (erfc) is defined as: [Equation 27] The erfc function is related to erf function by: [Equation 28] CS 515 Ibrahim Korpeoglu

Erf and Erfc functions The Q-function is related to erf and erfc functions by: [Equation 29] [Equation 30] [Equation 31] CS 515 Ibrahim Korpeoglu

Computation of probability that the received power is below/above a threshold
We said that Pr(d) is a random variable that is Gaussian distributed with mean m and std deviation s. Then: Probability that Pr(d) is above g is given by: Probability that Pr(d) is below g is given by: Pr(d) bar denotes the average (mean ) received power at d. Equation 32 Equation 33 CS 515 Ibrahim Korpeoglu

Percentage of Coverage Area
We are interested in the following problem Given a circular coverage area with radius R from a base station Given a desired threshold power level . Find out U(g), the percentage of useful service area i.e the percentage of area with a received signal that is equal or greater than g, given a known likelihood of coverage at the cell boundary CS 515 Ibrahim Korpeoglu

O is the origin of the cell O r: radial distance d from transmitter 0 <= r <= R r R Definition: P(Pr(r) > g) denotes probability that the random received power at a distance d = r is greater than threshold g within an incrementally small area dA Then U(g) can be found by the following integration over the area of the cell: Equation 34 CS 515 Ibrahim Korpeoglu

Integrating f(r) over Circle Area
Dq f(r) R B r D C Dr O CS 515 Ibrahim Korpeoglu

Using equation 32: Equation 33 The path loss at distance r can be expressed as: O d r R PL(from O to r) = PL(from O to d0) + PL(from d0 to R) - PL(from r to R) PL(from O to r) = PL(from O to d0) + PL(from d0 to R) + PL(from R to r) (O is the point where base station is located) Which can be formally expressed as: Equation 34 CS 515 Ibrahim Korpeoglu

Equation 33 can be expressed as follows using error function: Equation 35 By combining with Equation 34 Equation 36 CS 515 Ibrahim Korpeoglu

Let the following substitutions happen: Then Equation 37 Substitute t = a+ blog(r/R) Equation 38 CS 515 Ibrahim Korpeoglu

By choosing a signal level such that (i.e. a = ), we obtain: where Equation 39 The simplified formula above gives the percentage coverage assuming the mean received power at the cell boundary (r=R) is g. In other words, we are assuming: Prob(Pr(R) >= g) = 0.5 CS 515 Ibrahim Korpeoglu

Indoor and Outdoor Propagation

Outdoor Propagation We will look to the propagation from a transmitter in an outdoor environment The coverage area around a tranmitter is called a cell. Coverage area is defined as the area in which the path loss is at or below a given value. The shape of the cell is modeled as hexagon, but in real life it has much more irregular shapes. By playing with the antenna (tilting and changing the height), the size of the cell can be controlled. We will look to the propagation characteristics of the three outdoor environments Propagation in macrocells Propagation in microcells Propagation in street microcells CS 515 Ibrahim Korpeoglu

Macrocells Base stations at high-points Coverage of several kilometers
The average path loss in dB has normal distribution Avg path loss is result of many forward scattering over a great many of obstacles Each contributing a random multiplicative factor Converted to dB, this gives a sum of random variable Sum is normally distributed because of central limit theorem CS 515 Ibrahim Korpeoglu

Macrocells In early days, the models were based on emprical studies
Okumura did comprehesive measurements in 1968 and came up with a model. Discovered that a good model for path loss was a simple power law where the exponent n is a function of the frequency, antenna heights, etc. Valid for frequencies in: 100MHz – 1920 MHz for distances: 1km – 100km CS 515 Ibrahim Korpeoglu

Okumura Model L50(d)(dB) = LF(d)+ Amu(f,d) – G(hte) – G(hre) – GAREA
L50: 50th percentile (i.e., median) of path loss LF(d): free space propagation pathloss. Amu(f,d): median attenuation relative to free space Can be obtained from Okumura’s emprical plots shown in the book (Rappaport), page 151. G(hte): base station antenna heigh gain factor G(hre): mobile antenna height gain factor GAREA: gain due to type of environment G(hte) = 20log(hte/200) m > hte > 30m G(hre) = 10log(hre/3) hre <= 3m G(hre) = 20log(hre/3) m > hre > 3m hte: transmitter antenna height hre: receiver antenna height Equation 40 CS 515 Ibrahim Korpeoglu

Hata Model Valid from 150MHz to 1500MHz A standard formula
For urban areas the formula is: L50(urban,d)(dB) = logfc loghte – a(hre) (44.9 – 6.55loghte)logd where fc is the ferquency in MHz hte is effective transmitter antenna height in meters (30-200m) hre is effective receiver antenna height in meters (1-10m) d is T-R separation in km a(hre) is the correction factor for effective mobile antenna height which is a function of coverage area a(hre) = (1.1logfc – 0.7)hre – (1.56logfc – 0.8) dB for a small to medium sized city Equation 41 CS 515 Ibrahim Korpeoglu

Microcells Propagation differs significantly
Milder propagation characteristics Small multipath delay spread and shallow fading imply the feasibility of higher data-rate transmission Mostly used in crowded urban areas If transmitter antenna is lower than the surrounding building than the signals propagate along the streets: Street Microcells CS 515 Ibrahim Korpeoglu

Macrocells versus Microcells
Item Macrocell Microcell Cell Radius 1 to 20km 0.1 to 1km Tx Power 1 to 10W 0.1 to 1W Fading Rayleigh Nakgami-Rice RMS Delay Spread 0.1 to 10ms 10 to 100ns Max. Bit Rate 0.3 Mbps 1 Mbps CS 515 Ibrahim Korpeoglu

Street Microcells Most of the signal power propagates along the street. The sigals may reach with LOS paths if the receiver is along the same street with the transmitter The signals may reach via indirect propagation mechanisms if the receiver turns to another street. CS 515 Ibrahim Korpeoglu

Street Microcells D Building Blocks B C A A A B C D
Breakpoint received power (dB) received power (dB) A A B n=2 n=2 Breakpoint 15~20dB C n=4 n=4~8 D log (distance) log (distance) CS 515 Ibrahim Korpeoglu

Indoor Propagation Indoor channels are different from traditional mobile radio channels in two different ways: The distances covered are much smaller The variablity of the environment is much greater for a much smaller range of T-R separation distances. The propagation inside a building is influenced by: Layout of the building Construction materials Building type: sports arena, residential home, factory,... CS 515 Ibrahim Korpeoglu

Indoor Propagation Indoor propagation is domited by the same mechanisms as outdoor: reflection, scattering, diffraction. However, conditions are much more variable Doors/windows open or not The mounting place of antenna: desk, ceiling, etc. The level of floors Indoor channels are classified as Line-of-sight (LOS) Obstructed (OBS) with varying degrees of clutter. CS 515 Ibrahim Korpeoglu

Indoor Propagation Buiding types Residential homes in suburban areas
Residential homes in urban areas Traditional office buildings with fixed walls (hard partitions) Open plan buildings with movable wall panels (soft partitions) Factory buildings Grocery stores Retail stores Sport arenas CS 515 Ibrahim Korpeoglu

Indoor propagation events and parameters
Temporal fading for fixed and moving terminals Motion of people inside building causes Ricean Fading for the stationary receivers Portable receivers experience in general: Rayleigh fading for OBS propagation paths Ricean fading for LOS paths. Multipath Delay Spread Buildings with fewer metals and hard-partitions typically have small rms delay spreads: 30-60ns. Can support data rates excess of several Mbps without equalization Larger buildings with great amount of metal and open aisles may have rms delay spreads as large as 300ns. Can not support data rates more than a few hundred Kbps without equalization. Path Loss The following formula that we have seen earlier also describes the indoor path loss: PL(d)[dBm] = PL(d0) + 10nlog(d/d0) + Xs n and s depend on the type of the building Smaller value for s indicates the accuracy of the path loss model. CS 515 Ibrahim Korpeoglu

Path Loss Exponent and Standard Deviation Measured for Different Buildings
Frequency (MHz) n s (dB) Retail Stores 914 2.2 8.7 Grocery Store 1.8 5.2 Office, hard partition 1500 3.0 7.0 Office, soft partition 900 2.4 9.6 1900 2.6 14.1 Factory LOS Textile/Chemical 1300 2.0 4000 2.1 Paper/Cereals 6.0 Metalworking 1.6 5.8 Suburban Home Indoor Street Factory OBS 9.7 3.3 6.8 CS 515 Ibrahim Korpeoglu

In building path loss factors
Partition losses (same floor) Partition losses between floors Signal Penetration into Buildings CS 515 Ibrahim Korpeoglu

Partition Losses There are two kind of partition at the same floor:
Hard partions: the walls of the rooms Soft partitions: moveable partitions that does not span to the ceiling The path loss depends on the type of the partitions CS 515 Ibrahim Korpeoglu

Partition Losses Average signal loss measurements reported by various researches for radio paths obscructed by some common building material. Material Type Loss (dB) Frequency (MHz) All metal 26 815 Aluminim Siding 20.4 Concerete Block Wall 3.9 1300 Loss from one Floor 20-30 Turning an Angle in a Corridor 10-15 Concrete Floor 10 Dry Plywood (3/4in) – 1 sheet 1 9600 Wet Plywood (3/4in) – 1 sheet 19 Aluminum (1/8in) – 1 sheet 47 CS 515 Ibrahim Korpeoglu

Partition Losses between Floors
The losses between floors of a building are determined by External dimensions and materials of the building Type of construction used to create floors External surroundings Number of windows Presence of tinting on windows CS 515 Ibrahim Korpeoglu

Partition Losses between Floors
Average Floor Attenuation Factor in dB for One, Two, Three and Four Floors in Two Office Buildings Building FAF (dB) s (dB) Office Building 1 Through 1 Floor 12.9 7.0 Through 2 Floors 18.7 2.8 Through 3 Floors 24.4 1.7 Through 4 Floors 27.0 1.5 Office Building 2 16.2 2.9 27.5 5.4 31.6 7.2 CS 515 Ibrahim Korpeoglu

Signal Penetration Into Buildings
RF signals can penetrate from outside transmitter to the inside of buildings However the siganls are attenuated The path loss during penetration has been found to be a function of: Frequency of the signal The height of the building CS 515 Ibrahim Korpeoglu

Signal Penetration Into Buildings
Frequency (MHz) Loss (dB) 441 16.4 896.5 11.6 1400 7.6 Effect of Frequency Penetration loss decreases with increasing frequency Effect of Height Penetration loss decreases with the height of the building up-to some certain height At lower heights, the urban clutter induces greater attenuation and then it increases Shadowing affects of adjascent buildings CS 515 Ibrahim Korpeoglu

Conclusion More work needs to be done to understand the characteristics of wireless channels 3D numerical modeling approaches exist To achieve PCS, new and novel ways of classifying wireless environments will be needed that are both widely encompassing and reasonably compact. CS 515 Ibrahim Korpeoglu

Mobile Radio Propagation - Small-Scale Fading and Multipath
CS 515 Mobile and Wireless Networking Ibrahim Korpeoglu Computer Engineering Department Bilkent University

Road Map Intro to Impulse Response Model of a Multipath Channel
Intro to Small Scale Fading Doppler Shift Discrete-Time Signals Sinusiodal Functions Discrete-Time Systems and LTI Systems Exponential Representation of Sinus. Functions Filters Intro to Modulation Impulse Response Convolution (Discrete/ Continous) Discrete-time Impulse Response Model of Multipath Channel Power Delay Profile CS 515 © İbrahim Körpeoğlu, 2002

Small Scale Fading Describes rapid fluctuations of the amplitude, phase of multipath delays of a radio signal over short period of time or travel distance Caused by interference between two or more versions of the transmitted signal which arrive at the receiver at slightly different times. These waves are called multipath waves and combine at the receiver antenna to give a resultant signal which can vary widely in amplitude and phase CS 515 © İbrahim Körpeoğlu, 2002

Small Scale Multipath Propagation
Effects of multipath Rapid changes in the signal strength Over small travel distances, or Over small time intervals Random frequency modulation due to varying Doppler shifts on different multiples signals Time dispersion (echoes) caused by multipath propagation delays Multipath occurs because of Reflections Scattering CS 515 © İbrahim Körpeoğlu, 2002

Multipath At a receiver point
Radio waves generated from the same transmitted signal may come from different directions with different propagation delays with (possibly) different amplitudes (random) with (possibly) different phases (random) with different angles of arrival (random). These multipath components combine vectorially at the receiver antenna and cause the total signal to fade to distort CS 515 © İbrahim Körpeoğlu, 2002

Multipath Components Radio Signals Arriving from different directions to receiver Component 1 Component 2 Component N Receiver may be stationary or mobile. CS 515 © İbrahim Körpeoğlu, 2002

Mobility Other Objects in the radio channels may be mobile or stationary If other objects are stationary Motion is only due to mobile Fading is purely a spatial phenomenon (occurs only when the mobile receiver moves) The spatial variations as the mobile moves will be perceived as temporal variations Dt = Dd/v Fading may cause disruptions in the communication CS 515 © İbrahim Körpeoğlu, 2002

Factors Influencing Small Scale Fading
Multipath propagation Presence of reflecting objects and scatterers cause multiple versions of the signal to arrive at the receiver With different amplitudes and time delays Causes the total signal at receiver to fade or distort Speed of mobile Cause Doppler shift at each multipath component Causes random frequency modulation Speed of surrounding objects Causes time-varying Doppler shift on the multipath components CS 515 © İbrahim Körpeoğlu, 2002

Factors Influencing Small Scale Fading
Transmission bandwidth of the channel The transmitted radio signal bandwidth and bandwidth of the multipath channel affect the received signal properties: If amplitude fluctuates of not If the signal is distorted or not CS 515 © İbrahim Körpeoğlu, 2002

Doppler Effect Whe a transmitter or receiver is moving, the frequency of the received signal changes, i.e. İt is different than the frequency of transmissin. This is called Doppler Effect. The change in frequency is called Doppler Shift. It depends on The relative velocity of the receiver with respect to transmitter The frequenct (or wavelenth) of transmission The direction of traveling with respect to the direction of the arriving signal. CS 515 © İbrahim Körpeoğlu, 2002

Doppler Shift – Transmitter is moving
The frequency of the signal that is received behind the transmitter will be smaller The frequency of the signal that is received in front of the transmitter will be bigger CS 515 © İbrahim Körpeoğlu, 2002

Doppler Shift –Recever is moving
Dl X q Y d v A mobile receiver is traveling from point X to point Y CS 515 © İbrahim Körpeoğlu, 2002

Doppler Shift The Dopper shift is positive
If the mobile is moving toward the direction of arrival of the wave The Doppler shift is negative If the mobile is moving away from the direction of arrival of the wave. CS 515 © İbrahim Körpeoğlu, 2002

Impulse Response Model of a Multipath Channel
The wireless channel charcteristics can be expressed by impulse response function The channel is time varying channel when the receiver is moving. Lets assume first that time variation due strictly to the receiver motion (t = d/v) Since at any distance d = vt, the received power will ve combination of different incoming signals, the channel charactesitics or the impulse response funcion depends on the distance d between trandmitter and receiver CS 515 © İbrahim Körpeoğlu, 2002

Basic Digital Signal Processing (DSP) Concepts
Introduction to Signals and Systems

Signals and Systems Discrete-Time Signals Continuous Time Signals
Time-domain representation Transform domain representation Continuous Time Signals Transformations Fourier Transform Laplace Transform Z-transform Systems Linear Time Invariant (LTI) Systems Time Varying Systems Impulse Response of Systems CS 515 © İbrahim Körpeoğlu, 2002

Discrete-Time Signals – Time Domain Representation
Signals are represented as series of numbers called samples. A sample value is denoted with x[n]. (n is an integer). A discrete-time signal is represented by {x[n]}, which is also called a sequence. Example: {x[n]} = {…,-0.95, -0.2, 2.7, 1.1, -3.67, -0.7, 4.1, …} n=0 x[n] is called nth sample of the signal. X[n] could be a real number or complex number or an integer. If complex: {x[n]} = {Xre[n] + [Xim[n]} The signal may be finite length or infinite length. A sequence of samples can be extended by appending zeros. CS 515 © İbrahim Körpeoğlu, 2002

Discrete-Time Signals – Time Domain Representation
3 2 … … 1 n -1 -2 {x[n]} = {…, -2, 1, 3, 0, 2, 1, 0, -1, -2, -1, -2, 0, 2,…} CS 515 © İbrahim Körpeoğlu, 2002

Operations on Sequences
A single-input, single output system operates on a sequence, called the input sequence and develope an other sequence, called output sequence. In some system, there may be more than one input and/or more than one output. Basic Operations Let x[n] and y[n] be two input sequences: Modulation: w1[n] = x[n] · y[n] (product of the corresponding sample values) A device implementing modulation is called modulator. Addition: w2[n] = x[n] + y[n] (sum of the corresponding sample values) A device implementing addition is called adder. Multiplication: w3[n] = Ax[n] (multiply each sample by A) A device implementing multiplication is called multiplier. CS 515 © İbrahim Körpeoğlu, 2002

Operations on Sequences
4) Time-shifting: w4[n] = x[n-N], N is integer. If N > 0, then this is a delaying operation If N < 0, then this is a advancing operation The device implementing delaying operation by one sample is called unit delay, and a device implementing advancing operation by one sample is called unit advance. 5) Time-reversing (Folding): w6[n] = x[-n] 6) Pick-off node provides multiple copies of same sequence. CS 515 © İbrahim Körpeoğlu, 2002

Basic Operations on Sequences
x[n] X w1[n] x[n] + w2[n] Modulator Adder y[n] y[n] w1[n] = x[n]y[n] w2[n] = x[n]+y[n] A x[n] w3[n] x[n] z -1 w4[n] Multiplier Unit Delay w3[n] = Ax[n] w4[n] = x[n-1] x[n] x[n]] x[n] z w5[n] Unit Advance w5[n] = x[n+1] Pick-off Node x[n] CS 515 © İbrahim Körpeoğlu, 2002

Discrete-time System x[n-1] x[n-2] x[n-3] x[n] z-1 z-1 z-1 a1 a2 a3 a4
+ y[n] y[n] = a1x[n] + a2x[n-1] + a3x[n-2] + a4x[n-3] CS 515 © İbrahim Körpeoğlu, 2002

Periodic and Aperiodic Signal
A sequence is periodic if: A sequence is aperiodic if it is not periodic. Periodic sequences will be denoted with ~ on top of them. CS 515 © İbrahim Körpeoğlu, 2002

Energy and Power Signals
CS 515 © İbrahim Körpeoğlu, 2002

Basic Sequences Unit Sample Real Sinusoidal Sequences
Exponential Sequences Complex Exponentials Real and Imaginary parts Real Exponentials CS 515 © İbrahim Körpeoğlu, 2002

Unit Sample Sequence Also called as discrete-time impulse or the unit impulse. 1 The unit sample sequence shifted by k samples is given by: 1 k Any arbitrary sequence can be represented as the sum of weighted time-shifted unit sample sequences. Knowing the response of a LTI system to unit impulse, we can compute its response to any arbitrary input sequence. CS 515 © İbrahim Körpeoğlu, 2002

Real Sinusoidal Sequence

Example – x[n] = 1.5cosw0n w0=0.1p w0=0.2p w0=0.9p w0=0.9p CS 515
© İbrahim Körpeoğlu, 2002

Example – x[n] = 1.5cosw0n w0=1.1p w0=1.2p CS 515

Exponential Sequences

A complex exponential: real part

Real Exponentials X[n]=0.2(1.2)n a>1 Growing X[n]=20(0.9)n < 1
Decaying CS 515 © İbrahim Körpeoğlu, 2002

Discrete-Time Systems
Given an input sequence, generates an output sequence. If all signals are digital signals (which is the case for practical systems), then it is called a digital filter. Classification Linear System Shift-Invariant System Causal System Stable System Passive and Lossless Systems CS 515 © İbrahim Körpeoğlu, 2002

Linear System A system where superposition principle always holds.

Shift-Invariant System
If indices n is not related to specific time instants, then the system is called time invariant. This property ensures that for a specified input, the output of the system is independent of the time the input is applied. An LTI (Linear Time Invariant) System is a system that satisfies both linearity and time-invariance properties. CS 515 © İbrahim Körpeoğlu, 2002

Other System Properties
Causal System: In a causal discrete time system, the n0th output sample y[n0] depends only on input samples x[n] for n <= n0 and does not depend on input Samples for n > n0. Stable System: A discrete-time system is stable if and only if, for every bounded input, the output is also bounded. Passive and Lossless System: A disceret-time system is passive if, for every finite energy input sequence x[n], the output sequence y[n] has, at most, the same energy. A disceret-time system is lossless if, for every finite energy input sequence x[n], the output sequence y[n] has the same energy. CS 515 © İbrahim Körpeoğlu, 2002

Impulse Response The response of a digital filter to a unit sample sequence {d[n]} is called the unit response, or simply, the impulse response, and denoted by: {h[n]} A linear time invariant system (filter) is completely characterized in the time-domain by its impulse response. Example: A system is given with the equation below: y[n] = a1x[n] + a2x[n-1] + a3x[n-2] + a4x[n-3] By setting x[n] = d[n], we can find out the impulse response of the system: h[n] = a1d[n] + a2d[n-1] + a3d[n-2] + a4d[n-3] {h[n]} = {a1, a2, a3, a4} CS 515 © İbrahim Körpeoğlu, 2002

LTI: Input-Output Relationship
{x[n]} = {…,0, 0, 0.5, 0, 0, 1.5, -1, 0, 1, 0, 0.75, 0, 0, …} An arbitrary sequence x[n] in the time domain can be represented as a weighted sum of some basic sequences and its delayed versions. A commonly used basic sequence is unit sample sequence (unit impulse). Sequence x[n] described above can be expressed as: CS 515 © İbrahim Körpeoğlu, 2002

System Given d[n] h[n] (impulse response of the LTI system) Compute y[n] LTI System x[n] y[n] Lets compute the response of the this system (filter) to input x[n], where Since the system is time-invariant, the response to d[n-n0] will be h[n-n0]. Then: CS 515 © İbrahim Körpeoğlu, 2002

Any x[n] can be expressed as the weighted linear combination of delayed and advanced unit sample sequences as follows: x[k] denotes the kth sample value of {x[n]} By using the fact that the response of system to d[n-k] is h[n-k]. by change of variables is called convolution sum of x[n] and y[n] CS 515 © İbrahim Körpeoğlu, 2002

Example CS 515 © İbrahim Körpeoğlu, 2002

Unit Impulse and Response
d(n) h(n) CS 515 © İbrahim Körpeoğlu, 2002

Computation of Response
# k … -7 -6 -5 -4 -3 -2 -1 1 2 3 4 5 6 7 x[k] h[k] h[-k] h[1-k] h[2-k] h[3-k] h[4-k] 8 h[5-k] 9 h[6-k] 10 h[7-k] x To compute y[n], multiple row labeled x[k] with row labeled h[n-k]. Example: To compute y[3] : multiply row numbered 1 with row numbered 6. = (-2x-1)+(0x0)+(1x2)+(-1x1)+(3x0) = 3 CS 515 © İbrahim Körpeoğlu, 2002

Response CS 515 © İbrahim Körpeoğlu, 2002

Response Sequence: y[n]
x[n] y[n] {y[n]} = {…0,0,0,-2,-4,1,3,1,5,1,-3,0,0,0,…} The convolution of two finite-length sequences results in a finite-length sequence. CS 515 © İbrahim Körpeoğlu, 2002

Simple Interconnection of Systems
Cascade Connection: h1[n] h2[n] h2[n] h1[n] Parallel Connection h1[n] h1[n] CS 515 © İbrahim Körpeoğlu, 2002

Classifications of LTI Discrete-Time Systems
If h[n] is of finite length, such that h[n]=0 for n<N1 and n>N2, then the system is called finite impulse response (FIR) discrete time system. In this case, the convolution reduces to: If h[n] is of infinite length, then the system is called infinite impulse response (IIR) discrete-time system. For a causal system, the convolution is expressed as: CS 515 © İbrahim Körpeoğlu, 2002

Random Signals Deterministic Signal: They can be uniquely determined by a mathematical expression or a table lookup. Random Signal (Stochastic Signal): Each sample value is generated in a random fashion. A discrete time random signal consists of a typically infinite collection (or ensemble) of discrete time sequences {X[n]}. At any given time index n, the observed sample value x[n] is the value taken by the random variable X[n]. A random process is a family of random variables {X[n]}. CS 515 © İbrahim Körpeoğlu, 2002

Power of Discrete-Time Random Signals

Continuous Time Signals
Continues Time Impulse: The Unit Impulse d(t) is continues time signal that has the value ∞ at time t=0, and zero for all other time t. de(t) approximates d(t). δe(t) 1/ε The area under the spike is 1 t -ε/2 ε/2 CS 515 © İbrahim Körpeoğlu, 2002

Continuous Time Signals
An arbitrary signal x(t) can be approximated using staircase method: Superposition of impulses Shifting Property CS 515 © İbrahim Körpeoğlu, 2002

Impulse Response of Unit Impulse
LTI System d(t) h(t) d(t) is the unit impulse h(t) is called the impulse response of the system. We denote: By time invariance: CS 515 © İbrahim Körpeoğlu, 2002

Response of System to arbitrary continues time signal x(t)
Convolution Properties CS 515 © İbrahim Körpeoğlu, 2002

Multipath Channel Modeling
Impulse Response Model of a Multipath Wireless Channel

The wireless channel characteristics can be expressed by impulse response function The channel is time varying channel when the receiver is moving. Lets assume first that time variation due strictly to the receiver motion (t = d/v) Since at any distance d = vt, the received power will ve combination of different incoming signals, the channel charactesitics or the impulse response funcion depends on the distance d between trandmitter and receiver CS 515 © İbrahim Körpeoğlu, 2002

d = vt v d A receiver is moving along the ground at some constant velocity v. The multipath components that are received at the receiver will have different propagation delays depending on d: distance between transmitter and receiver. Hence the channel impulse response depends on d. Lets x(t) represents the transmitter signal y(d,t) represents the received signal at position d. h(d,t) represents the channel impulse response which is dependent on d (hence time-varying d=vt). CS 515 © İbrahim Körpeoğlu, 2002

Multipath Channel Model
Building Multipath Channel 2nd MC Base Station 1st MC Mobile 2 Building Building 1st MC 4th MC Multipath Channel 2nd MC Mobile 1 Building 3rd MC (Multipath Component) CS 515 © İbrahim Körpeoğlu, 2002

Wireless Multipath Channel h(d,t) x(t) y(t) The channel is linear time-varying channel, where the channel characteristics changes with distance (hence time, t = d/v) CS 515 © İbrahim Körpeoğlu, 2002

Impulse Response Model
We assume v is constant over short time. x(t): transmitted waveform y(t): received waveform h(t,): impulse response of the channel. Depends on d (and therefore t=d/v) and also to the multiple delay for the channel for a fixed value of t. t is the multipath delay of the channel for a fixed value of t. CS 515 © İbrahim Körpeoğlu, 2002

Introduction to Modulation
Baseband and Passband Signals and Comlex Envelopes (We will make a very brief introduction in order to better understand Multipath Channel Model)

Modulation Process of encoding information from a message source in a manner suitable for transmission. It involves translating a baseband message signal to a bandpass signal at frequencies that are very high compared to the base baseband frequency. Bandpass signal is called modulated signal Baseband signal is called modulating signal Modulation can be done by varying the amplitude, phase, or frequency of the carrier in accordance with the amplitude of the message signal. Assume we have a continues-time message signal m(t). Next we will show how we can modulate this signal to a carrier using AM modulation. AM Modulator m(t) sAM(t) CS 515 © İbrahim Körpeoğlu, 2002

Analog Modulation/Demodulation
Source Sink Wireless Channel Modulator Demodulator Baseband Signal with frequency fmesg (Modulating Signal) Bandpass Signal with frequency fc (Modulated Signal) Original Signal with frequency fmesg fc >> fmesg CS 515 © İbrahim Körpeoğlu, 2002

AM Modulation - Example
1/fmesg 1/fc CS 515 © İbrahim Körpeoğlu, 2002

Complex Envelope Modulation index k = peak_mesg_signal_amplitude / peak_carrier_amplitide The AM signal sAM(t) can also be expressed as: m(t) Accos(2pfct) g(t) is called the complex envelope AM signal CS 515 © İbrahim Körpeoğlu, 2002

Power of a bandpass signal
Let x(t) be a bandpass signal. Couch shows that average power of bandpass signal is: Lets have an intuition by an example CS 515 © İbrahim Körpeoğlu, 2002

...Continue with Multipath Channel Impulse Response Model

x(t) y(t) Bandpass Channel Impulse Response Model c(t) r(t) Baseband Equivalent Channel Impulse Response Model CS 515 © İbrahim Körpeoğlu, 2002

c(t) is the complex envelope representation of the transmitted signal r(t) is the complex envelope representation of the received signal hb(t,) is the complex baseband impulse response CS 515 © İbrahim Körpeoğlu, 2002

Discrete-time Impulse Response Model of Multipath Channel
Amplitude of Multipath Component There are N multipath components (0..N-1) o= 0 1= D Excess Delay Bin i= (i)D N-1= (N-1)D  (excess delay) D N-1 0 2 i Excess delay: relative delay of the ith multipath componentas compared to the first arriving component i : Excesss delay of ith multipath component, ND: Maximum excess delay CS 515 © İbrahim Körpeoğlu, 2002

Multipath Components arriving to a Receiver
Ignore the fact that multipath components arrive with different angles, and assume that they arriving with the same angle in 3D. 1 2 N-2 N-1 Nth Component t t0=0 t1 tN-3 tN-2 tN-1 (relative delay of multipath Comnponent) Each component will have different Amplitude (ai) and Phase (θi) CS 515 © İbrahim Körpeoğlu, 2002

Baseband impulse response of the Channel

Discrete-Time Impulse Response Model for a Multipath Channel
hb(t,) t t3 (t3) t2 (t2) t1 (t1) t0 (t0) o 1 2 3 4 5 6 N-2 N-1 CS 515 © İbrahim Körpeoğlu, 2002

Time-Invariance Assumption
If the channel impulse response is assumed to be time-invariant over small-scale time or distance interval, then the channel impulse response can be simplified as: When measuring or predicting hb(t), a probing pulse p(t) which approximates the unit impulse function is used at the transmitter. That is: This is called sounding the channel to determine impulse response. CS 515 © İbrahim Körpeoğlu, 2002

Complex Baseband Impulse Response
Baseband impulse response hb(t) is a complex number and therefore has a magnitude (amplitude) ai and a phase θi. hb(t) = aiejqi hb(t) ai hb(t) = ai(cosqi+jsinqi) qi |hb(t)| = ai you can think of it also as a vector that starts at origin. CS 515 © İbrahim Körpeoğlu, 2002

Amplitudes and Phases of Multipath Components
1st Arriving Multipath Component (Say 0th Component) q0=0 (phase) fc 2a0 Two components emerge from the same source at the same time. They belong to the same transmitter signal. But they travel different paths. They arrive at the same receiver with time difference equal to ti. fc 2ai ith Multipath Component qi=2pfcti t ti qi is expressed in radians CS 515 © İbrahim Körpeoğlu, 2002

Components arriving at the same time
What happens if two or more multipath components are with the same access delay bin (arrive at the same time)? Then the received signal is the vectorial addition of two multipath signals. R Example: Lets assume two signals S1 and S2 arrive at the same time at the receiver: a3 S1 a1 S2 q3 a2 q2 q1 R is the combined receiver signal. CS 515 © İbrahim Körpeoğlu, 2002

Components arriving at the same time
The amplitude and phase of the combined signal (R) depends on the amplitudes and phases of the two components. Depending on the values of the phases of the components, the combined affect may weaken or strengthen the amplitude of the combined signal. It is possible that the two signals may totally cancel each other depending on their relative phases on their amplitudes. CS 515 © İbrahim Körpeoğlu, 2002

Example 1 – Addition of Two Signals
MC: Multipath Component 1st MC 2st MC Combined Signal a1/a2=1 q1=p/16 q2 =p CS 515 © İbrahim Körpeoğlu, 2002

Example 2 – Addition of Two Signals
1st MC 2st MC Combined Signal a1/a2=1/3 q1=p/16 q2 =p CS 515 © İbrahim Körpeoğlu, 2002

Power Delay Profile For small-scale fading, the power delay profile of the channel is found by taking the spatial average of over a local area (small-scale area). If p(t) has a time duration much smaller than the impulse response of the multipath channel, the received power delay profile in a local area is given by: The bar represents the average over the local area of Gain k relates the transmitter power in the probing pulse p(t) to the total received power in a multipath delay profile. CS 515 © İbrahim Körpeoğlu, 2002

Example power delay profile
Taken from Dimitrios Mavrakis Homepage: CS 515 © İbrahim Körpeoğlu, 2002

Appendix: Basic Filter Concept (Simple Analog Filters)

Filters Filter: A device that transmits only part of the incident energy and may thereby change the spectral distribution of energy: (a) high-pass filters transmit energy above a certain frequency (pass higher frequencies); (b) low-pass filters transmit energy below a certain frequency (pass lower frequencies); (c) bandpass filters transmit energy of a certain bandwidth; (d) band-stop filters transmit energy outside a specific frequency band. CS 515 © İbrahim Körpeoğlu, 2002

Low-pass Filter R1 (k) A simple RC low-pass filter V(t) C1 (nF) Vout
Vin Cutoff (or turnower) freqeuncy: Voltage gain: Voltage Gain compares Vout with Vin CS 515 © İbrahim Körpeoğlu, 2002

Low-pass Filter Let f0=1000 Hertz Frequency f0 Voltage Gain = 0.707

Power and Voltage rms: root mean squate Vm : The maximum voltage
Vrms: The root mean square value of the time varying voltage P-bar: denotes the mean(average power) over a full cycle (periof) T: period of the sinosoidal wave f: frequency of the sinosoidal wave CS 515 © İbrahim Körpeoğlu, 2002

High-pass Filter C1 (nF) A simple RC high-pass filter Vin V(t) R1 (k)
Vout Cutoff (or turnower) freqeuncy: Voltage gain: CS 515 © İbrahim Körpeoğlu, 2002

High-pass Filter Let f0=1000 Hertz f0 Voltage Gain Frequency CS 515

Comparison Low, High and Bandpass Filters
Highpass filter Lowpass filter CS 515 © İbrahim Körpeoğlu, 2002

Appendix: Continues Time Sinusoidal Signals

Sinusoidal Signal X(t) T X(t)=Asin(2pft) Here f = 1Hz T = 1/f = 1 sec
X(t)=1*sin(2pt) t CS 515 © İbrahim Körpeoğlu, 2002

Sinusoidal Signal X(t) T X(t)=Asin(2pft) Here f = 3 Hz
T = 1/3 = .33 sec A = 5 X(t)=5*sin(2p3t) anges in radian t CS 515 © İbrahim Körpeoğlu, 2002

Sinusoidal Signal X(t) T X(t)=Acos(2pft) Here f = 1Hz T = 1/f = 1 sec
X(t)=1*cos(2pt) t CS 515 © İbrahim Körpeoğlu, 2002

Sinusoidal Signal – q = 0 radian
X(t) X(t)=Acos(2pft) Here f = 1Hz T = 1/f = 1 sec A = 2 X(t)=2*cos(2pt) CS 515 © İbrahim Körpeoğlu, 2002

Sinusoidal Signal - q = 1 radian
X(t) X(t)=Acos(2pft+q) Here f = 1Hz T = 1/f = 1 sec A = 2 = 1 radian (57 degrees) X(t)=2*cos(2pt+1) CS 515 © İbrahim Körpeoğlu, 2002

Sinusoidal Signal - q = 2 radians
X(t)=2*cos(2pt+2) q = 2 = 2*57 degrees CS 515 © İbrahim Körpeoğlu, 2002

Sinusoidal Signal - q = 3 radians
X(t)=2*cos(2pt+3) q = 3 radians = 3*57 degrees CS 515 © İbrahim Körpeoğlu, 2002

Appendix: Complex or Exponential Form Representation of Sinusoidal Signals

Complex numbers and Exponential Form
Imaginary W (complex number) jb M W = a + jb W = M(cosq + jsinq) q a By Euler’s Theorem: cosq + jsinq = ejq Therfore W = Mejq (Exponential or polar form) CS 515 © İbrahim Körpeoğlu, 2002

Sinusoids and Exponential Form Representation
Imaginary Axis W (complex number) jb M wt+q W = a + jb W = M(cosq + jsinq) wt (wt) is the angular distance that is traveled by the line M q Real Axis a Assume line rotates at an angular velocity w Then W is a complex function of time. W(t) = Mej(wt+q) Projection of this line on the real and imaginary axes: Wreal = M cos (wt + q), Wimag = M sin (wt + q) We will use the real component to represent a sinusoidal signal CS 515 © İbrahim Körpeoğlu, 2002

Exponential Form Representation
Let x(t) be a sinusoidal function as follows: We will represent x(t) in complex or exponential form as follows: A(t) is called complex envelope CS 515 © İbrahim Körpeoğlu, 2002

Mobile Radio Propagation - Small-Scale Fading and Multipath
CS 515 Mobile and Wireless Networking Fall 2002 İbrahim Körpeoğlu Computer Engineering Department Bilkent University

Relationship between Bandwidth and Receiver Power
What happens when two different signals with different bandwidths are sent through the channel? What is the receiver power characteristics for both signals? We mean the bandwith of the baseband signal The bandwidth of the baseband is signal is inversely related with its symbol rate. One symbol CS 515 © İbrahim Körpeoğlu

Bandwidth of Baseband Signals
Highbandwidth (Wideband) Signal Lowbandwidth (Narrowband) Signal Continuous Wave (CW) Signal t CS 515 © İbrahim Körpeoğlu

A pulsed probing signal (wideband)
Tbb Transmitter p(t) x(t): transmitted signal TREP x(t) Multipath Wireless Channel y(t) p(t) Multipath Wireless Channel r(t) Bandpass signals Baseband signals CS 515 © İbrahim Körpeoğlu

Received Power of Wideband Sİgnals
p(t) Multipath Wireless Channel r(t) The output r(t) will approximate the channel impulse response since p(t) approximates unit impulses. Assume the multipath components have random amplitudes and phases at time t. CS 515 © İbrahim Körpeoğlu

Received Power of Wideband Sİgnals
This shows that if all the multipath components of a transmitted signal is resolved at the receiver then: The average small scale received power is simply the sum of received powers in each multipath component. In practice, the amplitudes of individual multipath components do not fluctuate widely in a local area (for distance in the order of wavelength or fraction of wavelength). This means the average received power of a wideband signal do not fluctuate significantly when the receiver is moving in a local area. CS 515 © İbrahim Körpeoğlu

Received Power of Narrowband Sİgnals
Transmitter A CW Signal x(t): transmitted signal c(t) Assume now A CW signal transmitted into the same channel. Let comlex envelope will be: The instantaneous complex envelope of the received signal will be: The instantaneous power will be: CS 515 © İbrahim Körpeoğlu

Received Power of Narrowband Sİgnals
Over a local area (over small distance – wavelengths), the amplitude a multipath component may not change signicantly, but the phase may change a lot. For example: - if receiver moves l meters then phase change is 2p. In this case the component may add up posively to the total sum S. - if receiver moves l/4 meters then phase change is p/2 (90 degrees) . In this case the component may add up negatively to the total sum S, hence the instantaneous receiver power. Therefore for a CW (continues wave, narrowband) signal, the small movements may cause large fluctuations on the instantenous receiver power, which typifies small scale fading for CW signals. CS 515 © İbrahim Körpeoğlu

Wideband versus Narrowband Baseband Signals
However, the average received power for a CW signal over a local area is equivalent to the average received power for a wideband signal on the local area. This occurs because the phases of multipath components at different locations over the small-scale region are independently distributed (IID uniform) over [0,2p]. In summary: Received power for CW signals undergoes rapid fades over small distances Received power for wideband signals changes very little of small distances. However, the local area average of both signals are nearly identical. CS 515 © İbrahim Körpeoğlu

Small-Scale Multipath Measurements
Several Methods Direct RF Pulse System Spread Spectrum Sliding Correlator Channel Sounding Frequency Domain Channel Sounding These techniques are also called channel sounding techniques CS 515 © İbrahim Körpeoğlu

Direct RF Pulse System Tx fc Pulse Generator RF Link Rx Digital BPF
Oscilloscope BPF Detector CS 515 © İbrahim Körpeoğlu

Parameters of Mobile Multipath Channels
Time Dispersion Parameters Grossly quantifies the multipath channel Determined from Power Delay Profile Parameters include Mean Access Delay RMS Delay Spread Excess Delay Spread (X dB) Coherence Bandwidth Doppler Spread and Coherence Time CS 515 © İbrahim Körpeoğlu

Measuring PDPs Power Delay Profiles
Are measured by channel sounding techniques Plots of relative received power as a function of excess delay They are found by averaging intantenous power delay measurements over a local area Local area: no greater than 6m outdoor Local area: no greater than 2m indoor Samples taken at l/4 meters approximately For 450MHz – 6 GHz frequency range. CS 515 © İbrahim Körpeoğlu

Timer Dispersion Parameters
Determined from a power delay profile. Mean excess delay( ): Rms delay spread (st): CS 515 © İbrahim Körpeoğlu

Timer Dispersion Parameters
Maximum Excess Delay (X dB): Defined as the time delay value after which the multipath energy falls to X dB below the maximum multipath energy (not necesarily belonging to the first arriving component). It is also called excess delay spread. CS 515 © İbrahim Körpeoğlu

RMS Delay Spread CS 515 © İbrahim Körpeoğlu

PDP Outdoor CS 515 © İbrahim Körpeoğlu

PDP Indoor CS 515 © İbrahim Körpeoğlu

Noise Threshold The values of time dispersion parameters also depend on the noise threshold (the level of power below which the signal is considered as noise). If noise threshold is set too low, then the noise will be processed as multipath and thus causing the parameters to be higher. CS 515 © İbrahim Körpeoğlu

Coherence Bandwidth (BC)
Range of frequencies over which the channel can be considered flat (i.e. channel passes all spectral components with equal gain and linear phase). It is a definition that depends on RMS Delay Spread. Two sinusoids with frequency separation greater than Bc are affected quite differently by the channel. f1 Receiver f2 Multipath Channel Frequency Separation: |f1-f2| CS 515 © İbrahim Körpeoğlu

Coherence Bandwidth Frequency correlation between two sinusoids: 0 <= Cr1, r2 <= 1. If we define Coherence Bandwidth (BC) as the range of frequencies over which the frequency correlation is above 0.9, then s is rms delay spread. If we define Coherence Bandwidth as the range of frequencies over which the frequency correlation is above 0.5, then This is called 50% coherence bandwidth. CS 515 © İbrahim Körpeoğlu

Coherence Bandwidth Example:
For a multipath channel, s is given as 1.37ms. The 50% coherence bandwidth is given as: 1/5s = 146kHz. This means that, for a good transmission from a transmitter to a receiver, the range of transmission frequency (channel bandwidth) should not exceed 146kHz, so that all frequencies in this band experience the same channel characteristics. Equalizers are needed in order to use transmission frequencies that are separated larger than this value. This coherence bandwidth is enough for an AMPS channel (30kHz band needed for a channel), but is not enough for a GSM channel (200kHz needed per channel). CS 515 © İbrahim Körpeoğlu

Coherence Time Delay spread and Coherence bandwidth describe the time dispersive nature of the channel in a local area. They don’t offer information about the time varying nature of the channel caused by relative motion of transmitter and receiver. Doppler Spread and Coherence time are parameters which describe the time varying nature of the channel in a small-scale region. CS 515 © İbrahim Körpeoğlu

Doppler Spread Measure of spectral broadening caused by motion
We know how to compute Doppler shift: fd Doppler spread, BD, is defined as the maximum Doppler shift: fm = v/l If the baseband signal bandwidth is much greater than BD then effect of Doppler spread is negligible at the receiver. CS 515 © İbrahim Körpeoğlu

Coherence Time Coherence time is the time duration over which the channel impulse response is essentially invariant. If the symbol period of the baseband signal (reciprocal of the baseband signal bandwidth) is greater the coherence time, than the signal will distort, since channel will change during the transmission of the signal . TS Coherence time (TC) is defined as: TC f2 f1 Dt=t2 - t1 t1 t2 CS 515 © İbrahim Körpeoğlu

Coherence Time Coherence time is also defined as:
Coherence time definition implies that two signals arriving with a time separation greater than TC are affected differently by the channel. CS 515 © İbrahim Körpeoğlu

Types of Small-scale Fading
CS 515 © İbrahim Körpeoğlu

Flat Fading Occurs when the amplitude of the received signal changes with time For example according to Rayleigh Distribution Occurs when symbol period of the transmitted signal is much larger than the Delay Spread of the channel Bandwidth of the applied signal is narrow. May cause deep fades. Increase the transmit power to combat this situation. CS 515 © İbrahim Körpeoğlu

Flat Fading h(t,t) s(t) r(t) t << TS TS t TS+t Occurs when:
TS t TS+t Occurs when: BS << BC and TS >> st BC: Coherence bandwidth BS: Signal bandwidth TS: Symbol period st: Delay Spread CS 515 © İbrahim Körpeoğlu

Frequency Selective Fading
Occurs when channel multipath delay spread is greater than the symbol period. Symbols face time dispersion Channel induces Intersymbol Interference (ISI) Bandwidth of the signal s(t) is wider than the channel impulse response. CS 515 © İbrahim Körpeoğlu

Frequency Selective Fading
h(t,t) s(t) r(t) t >> TS TS t TS TS+t Causes distortion of the received baseband signal Causes Inter-Symbol Interference (ISI) Occurs when: BS > BC and TS < st As a rule of thumb: TS < st CS 515 © İbrahim Körpeoğlu

Fast Fading Due to Doppler Spread
Rate of change of the channel characteristics is larger than the Rate of change of the transmitted signal The channel changes during a symbol period. The channel changes because of receiver motion. Coherence time of the channel is smaller than the symbol period of the transmitter signal Occurs when: BS < BD and TS > TC BS: Bandwidth of the signal BD: Doppler Spread TS: Symbol Period TC: Coherence Bandwidth CS 515 © İbrahim Körpeoğlu

Slow Fading Due to Doppler Spread
Rate of change of the channel characteristics is much smaller than the Rate of change of the transmitted signal Occurs when: BS >> BD and TS << TC BS: Bandwidth of the signal BD: Doppler Spread TS: Symbol Period TC: Coherence Bandwidth CS 515 © İbrahim Körpeoğlu

Different Types of Fading
TS Flat Fast Fading Flat Slow Fading Symbol Period of Transmitting Signal st Frequency Selective Slow Fading Frequency Selective Fast Fading TC TS Transmitted Symbol Period With Respect To SYMBOL PERIOD CS 515 © İbrahim Körpeoğlu

Different Types of Fading
BS Frequency Selective Fast Fading Frequency Selective Slow Fading Transmitted Baseband Signal Bandwidth BC Flat Fast Fading Flat Slow Fading BD BS Transmitted Baseband Signal Bandwidth With Respect To BASEBAND SIGNAL BANDWIDTH CS 515 © İbrahim Körpeoğlu

Fading Distributions Describes how the received signal amplitude changes with time. Remember that the received signal is combination of multiple signals arriving from different directions, phases and amplitudes. With the received signal we mean the baseband signal, namely the envelope of the received signal (i.e. r(t)). Its is a statistical characterization of the multipath fading. Two distributions Rayleigh Fading Ricean Fading CS 515 © İbrahim Körpeoğlu

Rayleigh and Ricean Distributions
Describes the received signal envelope distribution for channels, where all the components are non-LOS: i.e. there is no line-of–sight (LOS) component. Describes the received signal envelope distribution for channels where one of the multipath components is LOS component. i.e. there is one LOS component. CS 515 © İbrahim Körpeoğlu

Rayleigh Fading CS 515 © İbrahim Körpeoğlu

Rayleigh Rayleigh distribution has the probability density function (PDF) given by: s2 is the time average power of the received signal before envelope detection. s is the rms value of the received voltage signal before envelope detection Remember: (see end of slides 5) CS 515 © İbrahim Körpeoğlu

Rayleigh The probability that the envelope of the received signal does not exceed a specified value of R is given by the CDF: CS 515 © İbrahim Körpeoğlu

Rayleigh PDF 0.6065/s mean = 1.2533s median = 1.177s
variance = s2 s 2s 3s 4s 5s CS 515 © İbrahim Körpeoğlu

Ricean Distribution When there is a stationary (non-fading) LOS signal present, then the envelope distribution is Ricean. The Ricean distribution degenerates to Rayleigh when the dominant component fades away. CS 515 © İbrahim Körpeoğlu

Level Crossing Rate (LCR)
Threshold (R) LCR is defined as the expected rate at which the Rayleigh fading envelope, normalized to the local rms signal level, crosses a specified threshold level R in a positive going direction. It is given by: CS 515 © İbrahim Körpeoğlu

Average Fade Duration Defined as the average period of time for which the received signal is below a specified level R. For Rayleigh distributed fading signal, it is given by: CS 515 © İbrahim Körpeoğlu

Fading Model – Gilbert-Elliot Model
Fade Period Signal Amplitude Threshold Time t Good (Non-fade) Bad (Fade) CS 515 © İbrahim Körpeoğlu

Gilbert-Elliot Model 1/AFD Good Bad 1/ANFD
(Non-fade) Bad (Fade) 1/ANFD The channel is modeled as a Two-State Markov Chain. Each state duration is memory-less and exponentially distributed. The rate going from Good to Bad state is: 1/AFD (AFD: Avg Fade Duration) The rate going from Bad to Good state is: 1/ANFD (ANFD: Avg Non-Fade Duration) CS 515 © İbrahim Körpeoğlu

Modulation Techniques for Mobile Radio
CS 515 Mobile and Wireless Networking Ibrahim Korpeoglu Computer Engineering Department Bilkent University CS 515 © Ibrahim Korpeoglu

What is modulation Modulation is the process of encoding information from a message source in a manner suitable for transmission It involves translating a baseband message signal to a bandpass signal at frequencies that are very high compared to the baseband frequency. Baseband signal is called modulating signal Bandpass signal is called modulated signal CS 515 © Ibrahim Korpeoglu

Modulation Techniques
Modulation can be done by varying the Amplitude Phase, or Frequency of a high frequency carrier in accordance with the amplitude of the message signal. Demodulation is the inverse operation: extracting the baseband message from the carrier so that it may be processed at the receiver. CS 515 © Ibrahim Korpeoglu

Analog/Digital Modulation
Analog Modulation The input is continues signal Used in first generation mobile radio systems such as AMPS in USA. Digital Modulation The input is time sequence of symbols or pulses. Are used in current and future mobile radio systems CS 515 © Ibrahim Korpeoglu

Goal of Modulation Techniques
Modulation is difficult task given the hostile mobile radio channels Small-scale fading and multipath conditions. The goal of a modulation scheme is: Transport the message signal through the radio channel with best possible quality Occupy least amount of radio (RF) spectrum. CS 515 © Ibrahim Korpeoglu

Frequency versus Amplitude Modulation
Frequency Modulation (FM) Most popular analog modulation technique Amplitude of the carrier signal is kept constant (constant envelope signal), the frequency of carrier is changed according to the amplitude of the modulating message signal; Hence info is carried in the phase or frequency of the carrier. Has better noise immunity: atmospheric or impulse noise cause rapid fluctuations in the amplitude of the received signal Performs better in multipath environment Small-scale fading cause amplitude fluctuations as we have seen earlier. Can trade bandwidth occupancy for improved noise performance. Increasing the bandwith occupied increases the SNR ratio. The relationship between received power and quality is non-linear. Rapid increase in quality for an increase in received power. Resistant to co-channel interference (capture effect). CS 515 © Ibrahim Korpeoglu

Amplitude Modulation (AM)
Changes the amplitude of the carrier signal according to the amplitude of the message signal All info is carried in the amplitude of the carrier There is a linear relationship between the received signal quality and received signal power. AM systems usually occupy less bandwidth then FM systems. AM carrier signal has time-varying envelope. CS 515 © Ibrahim Korpeoglu

Amplitude Modulation The amplitude of high-carrier signal is varied according to the instantaneous amplitude of the modulating message signal m(t). AM Modulator m(t) sAM(t) CS 515 © Ibrahim Korpeoglu

Modulation Index of AM Signal
For a sinusoidal message signal Index is defined as: SAM(t) can also be expressed as: g(t) is called the complex envelope of AM signal. CS 515 © Ibrahim Korpeoglu

AM Modulation/Demodulation
Source Sink Wireless Channel Modulator Demodulator Baseband Signal with frequency fm (Modulating Signal) Bandpass Signal with frequency fc (Modulated Signal) Original Signal with frequency fm fc >> fm CS 515 © Ibrahim Korpeoglu

AM Modulation - Example
1/fmesg 1/fc CS 515 © Ibrahim Korpeoglu

Angle Modulation Angle of the carrier is varied according to the amplitude of the modulating baseband signal. Two classes of angle modulation techniques: Frequency Modulation Instantanoues frequency of the carrier signal is varied linearly with message signal m(t) Phase Modulation The phase q(t) of the carrier signal is varied linearly with the message signal m(t). CS 515 © Ibrahim Korpeoglu

Angle Modulation FREQUENCY MODULATION
kf is the frequency deviation constant (kHz/V) If modulation signal is a sinusoid of amplitude Am, frequency fm: PHASE MODULATION kq is the phase deviation constant CS 515 © Ibrahim Korpeoglu

FM Example: 4 - + - - + -4 0.5 1 1.5 2 Message signal FM Signal
- + - - + -4 0.5 1 1.5 2 Message signal FM Signal Carrier Signal CS 515 © Ibrahim Korpeoglu

FM Index W: the maximum bandwidth of the modulating signal
Df: peak frequency deviation of the transmitter. Am: peak value of the modulating signal Example: Given m(t) = 4cos(2p4x103t) as the message signal and a frequency deviation constant gain (kf) of 10kHz/V; Compute the peak frequency deviation and modulation index! Answer: fm=4kHz Df = 10kHz/V * 4V = 40kHz. bf = 40kHz / 4kHz = 10 CS 515 © Ibrahim Korpeoglu

Spectra and Bandwidth of FM Signals
An FM Signal has 98% of the total transmitted power in a RF bandwidth BT Carson’s Rule Upper bound Lower bound Example: Analog AMPS FM system uses modulation index of Bf = 3 and fm = 4kHz. Using Carson’s Rule: AMPS has 32kHz upper bound and 24kHz lower bound on required channnel bandwidth. CS 515 © Ibrahim Korpeoglu

FM Demodulator Convert from the frequency of the carrier signal to the amplitude of the message signal FM Detection Techniques Slope Detection Zero-crossing detection Phase-locked discrimination Quadrature detection CS 515 © Ibrahim Korpeoglu

Proportional to the priginal Message Signal
Slope Detector Vin(t) Vout(t) Limiter V1(t) Differentiator V2(t) Envelope Detector Proportional to the priginal Message Signal CS 515 © Ibrahim Korpeoglu

Digital Modulation The input is discrete signals
Time sequence of pulses or symbols Offers many advantages Robustness to channel impairments Easier multiplexing of variuous sources of information: voice, data, video. Can accommodate digital error-control codes Enables encryption of the transferred signals More secure link CS 515 © Ibrahim Korpeoglu

Digital Modulation The modulating signal is respresented as a time-sequence of symbols or pulses. Each symbol has m finite states: That means each symbol carries n bits of information where n = log2m bits/symbol. ... Modulator T One symbol (has m states – voltage levels) (represents n = log2m bits of information) CS 515 © Ibrahim Korpeoglu

Factors that Influence Choice of Digital Modulation Techniques
A desired modulation scheme Provides low bit-error rates at low SNRs Power efficiecny Performs well in multipath and fading conditions Occupies minimum RF channel bandwidth Bandwidth efficieny Is easy and cost-effective to implement Depending on the demands of a particular system or application, tradeoffs are made when selecting a digital modulation scheme. CS 515 © Ibrahim Korpeoglu

Power Efficiency of Modulation
Power efficiency is the ability of the modulation technique to preserve fidelity of the message at low power levels. Usually in order to obtain good fidelity, the signal power needs to be increased. Tradeoff between fidelity and signal power Power efficiency describes how efficient this tradeoff is made Eb: signal energy per bit N0: noise power spectral density PER: probability of error CS 515 © Ibrahim Korpeoglu

Bandwidth Efficiency of Modulation
Ability of a modulation scheme to accommodate data within a limited bandwidth. Bandwidth efficiency reflect how efficiently the allocated bandwidth is utilized R: the data rate (bps) B: bandwidth occupied by the modulated RF signal CS 515 © Ibrahim Korpeoglu

Shannon’s Bound There is a fundamental upper bound on achievable bandwidth efficiency. Shannon’s theorem gives the relationship between the channel bandwidth and the maximum data rate that can be transmitted over this channel considering also the noise present in the channel. Shannon’s Theorem C: channel capacity (maximum data-rate) (bps) B: RF bandwidth S/N: signal-to-noise ratio (no unit) CS 515 © Ibrahim Korpeoglu

Tradeoff between BW Efficiency and Power Efficiency
There is a tradeoff between bandwidth efficiency and power efficiency Adding error control codes Improves the power efficiency Reduces the requires received power for a particular bit error rate Decreases the bandwidth efficiency Increases the bandwidth occupancy M-ary keying modulation Increases the bandwidth efficiency Decreases the power efficiency More power is requires at the receiver CS 515 © Ibrahim Korpeoglu

Example: SNR for a wireless channel is 30dB and RF bandwidth is 200kHz. Compute the theoretical maximum data rate that can be transmitted over this channel? Answer: CS 515 © Ibrahim Korpeoglu

Noiseless Channels and Nyquist Theorem
For a noiseless channel, Nyquist theorem gives the relationship between the channel bandwidth and maximum data rate that can be transmitted over this channel. Nyquist Theorem C: channel capacity (bps) B: RF bandwidth m: number of finite states in a symbol of transmitted signal Example: A noiseless channel with 3kHz bandwidth can only transmit maximum of 6Kbps if the symbols are binary symbols. CS 515 © Ibrahim Korpeoglu

Power Spectral Density of Digital Signals and Bandwith
What does signal bandwidth mean? Answer is based on Power Spectral Density (PSD) of Signals For a random signal w(t), PSD is defined as: CS 515 © Ibrahim Korpeoglu

PSD CS 515 © Ibrahim Korpeoglu

Fourier Analysis Joseph Fourier has shown that any periodic function F(f) with period T, can be constructed by summing a (possibly infinite) number of sines and cosines. Such a decomposition is called Fourier series and the coefficients are called the Fourier coefficients. A line graph of the amplitudes of the Fourier series components can be drawn as a function of frequency. Such a graph is called a spectrum or frequency spectrum. f0 is called the fundemantal frequency. The nth term is called nth harmonic. The coefficients of the nth harmonic are an and bn. CS 515 © Ibrahim Korpeoglu

Fourier Analysis The coefficients can be obtained from the periodic function F(t) as follows: CS 515 © Ibrahim Korpeoglu

Example: A Periodic Function
Find the fourier series of the periodic function f(x), where One period of f(x) is defined as: f(x) = x, -p < x < p T=2p p -p p 2p -p CS 515 © Ibrahim Korpeoglu

Example: Its Fourier Approximation
Domain: [-p, p] 1 harmonic 2 harmonics 3 harmonics 4 harmonics CS 515 © Ibrahim Korpeoglu

Example: Frequency Spectrum
For First 10 harmonics Each harmonic corresponds to a frequency that is multiple of the fundamental frequency CS 515 © Ibrahim Korpeoglu

Complex Form of Fourier Series
By substituting Euler’s Expresssion into Fouries expansion: It can be shown that the following is true: cn are the complex Fourier coefficients CS 515 © Ibrahim Korpeoglu

Digital Modulation - Continues
Line Coding Base-band signals are represented as line codes 1 1 1 1 Tb V Unipolar NRZ Tb V Bipolar RZ -V V Manchester NRZ -V Tb CS 515 © Ibrahim Korpeoglu

Geometric Representation of Modulation Signal
Digital Modulation involves Choosing a particular signal waveform for transmission for a particular symbol or signal For M possible signals, the set of all signal waveforms are: For binary modulation, each bit is mapped to a signal from a set of signal set S that has two signals We can view the elements of S as points in vector space CS 515 © Ibrahim Korpeoglu

Geometric Representation of Modulation Signal
Vector space We can represented the elements of S as linear combination of basis signals. The number of basis signals are the dimension of the vector space. Basis signals are orthogonal to each-other. Each basis is normalized to have unit energy: CS 515 © Ibrahim Korpeoglu

Example Two signal waveforms to be used for transmission
The basis signal Q I Constellation Diagram Dimension = 1 CS 515 © Ibrahim Korpeoglu

Constellation Diagram
Properties of Modulation Scheme can be inferred from Constellation Diagram Bandwidth occupied by the modulation increases as the dimension of the modulated signal increases Bandwidth occupied by the modulation decreases as the signal_points per dimension increases (getting more dense) Probability of bit error is proportional to the distance between the closest points in the constellation. Bit error decreases as the distance increases (sparse). CS 515 © Ibrahim Korpeoglu

Linear Modulation Techniques
Classify digital modulation techniques as: Linear The amplitude of the transmitted signal varies linearly with the modulating digital signal, m(t). They usually do not have constant envelope. More spectral efficient. Poor power efficiency Example: QPSK. Non-linear CS 515 © Ibrahim Korpeoglu

Binary Phase Shift Keying
Use alternative sine wave phase to encode bits Phases are separated by 180 degrees. Simple to implement, inefficient use of bandwidth. Very robust, used extensively in satellite communication. Q State 1 State CS 515 © Ibrahim Korpeoglu

BPSK Example 1 1 0 1 0 1 Data Carrier Carrier+ p BPSK waveform CS 515
Data Carrier Carrier+ p BPSK waveform CS 515 © Ibrahim Korpeoglu

Quadrature Phase Shift Keying
Multilevel Modulation Technique: 2 bits per symbol More spectrally efficient, more complex receiver. Two times more bandwidth efficient than BPSK Q 11 State 01 State 00 State 10 State Phase of Carrier: p/4, 2p/4, 5p/4, 7p/4 CS 515 © Ibrahim Korpeoglu

4 different waveforms cos+sin -cos+sin 11 01 00 10 cos-sin -cos-sin
CS 515 © Ibrahim Korpeoglu

Constant Envelope Modulation
Amplitude of the carrier is constant, regardless of the variation in the modulating signal Better immunity to fluctuations due to fading. Better random noise immunity Power efficient They occupy larger bandwidth CS 515 © Ibrahim Korpeoglu

Frequency Shift Keying (FSK)
The frequency of the carrier is changed according to the message state (high (1) or low (0)). Continues FSK Integral of m(x) is continues. CS 515 © Ibrahim Korpeoglu

FSK Example Data FSK Signal CS 515 © Ibrahim Korpeoglu

Medium Access Control for Wireless Links
CS 515 Mobile and Wireless Networking Ibrahim Korpeoglu Computer Engineering Department Bilkent University

Outline Random MAC Schemes
What we have see so far? PHY layer functions and parameters General Wireless System Architecture Media Access Control Classes of MAC protocols Simplex and Duplex Channels Coordinated MAC Schemes FDMA TDMA Capacity of TDMA systems and which factors affect the capacity. Spread Spectrum Access Methods FHMA Case study: Bluetooth CDMA Hybrid Spread Spectrum Schemes. Random MAC Schemes CSMA MACA and MACAW Case Study: IEEE MAC CS 515 © Ibrahim Korpeoglu

What we have seen so far? Physical layer functions
Get stream of bits and transport them to the other end. Modulation/Demodulation We have seen that this is not an easy task Large-scale path loss and Small-scale fading and multipath effects causes the received power at the receiver to Fluctuate (hard to decode the symbols (bits)) To decrease (Affects of interfering sources increases) Received signal power level affect the quality of the signals (information) that is transported. Received signal power level defines the Signal-to-Noise (SNR) ratio CS 515 © Ibrahim Korpeoglu

What we have seen so far? We have seen
That SNR and bandwidth of a channel affects Datarate – bps) of the wireless channel by Shannon limit The bit error rate (BER) on the channel. That multipath fading results in a wireless channel error model that changes states between good (low-error rate) and bad (high error-rate) Large-scale path loss defines the range of stations for different environments (LOS, urban,…) The above factors are important channel characteristics that affect the design of wireless systems architectures and design of the protocols and applications for wireless links/networks In short, we have seen so far some of Fundamental Concepts of Wireless Communication. CS 515 © Ibrahim Korpeoglu

What we will do now? We will look now to the protocols, algorithms, schemes that are developed over this wireless channels. How can we share a wireless channel: Results in Wireless Media Access Control Protocols How we can change base stations: Results in Handoff algorithms and protocols How can we seamlessly support mobile applications over wireless links: Results in mobility protocols like Mobile IP, Cellular IP, etc. How can we design efficient transport protocols over wireless links: Results in solutions like SNOOP, I-TCP, M-TCP, etc. How different wireless networks/systems are designed? Bluetooth, IEEE , GSM, etc. CS 515 © Ibrahim Korpeoglu

Wireless System Architecture and Functions
Applications TCP/IP Neighbor Discovery and Registration, Multicasting, Power Saving Modes, Address Translation (IP-MAC), Routing, Quality of Services, Subnet Security Wireless Subnet Controller Wireless Link Layer (Layers 1 and 2 in ISO/OSI Network Reference Model) Link Controller Medium Access Control, MAC level Scheduling, Link Layer Queueing, Link Layer Reliability – ACKs, NACKs, …. Transceiver Frame Controller Framing and frame synchronization, error control, CRC, bit scrambling, widening, …. Carrier frequency, channel bandwidth, carrier detect, Captude detect, channel data rate, modulation, Received signal strength (RSSI), transmit power, Power control, … Physical Radio CS 515 © Ibrahim Korpeoglu

Medium Access Control Wireless spectrum (frequency band) is a very precious and limited resource. We need to use this resource very efficiently We also want our wireless system to have high user capacity A lot of (multiple) users should be able to use the system at the same time. For these reasons most of the time, multiple users (or stations, computers, devices) need to share the wireless channel that is allocated and used by a system. The algorithms and protocols that enables this sharing by multiple users and controls/coordinates the access to the wireless channel (medium) from different users are called MEDIUM ACCESS, or MEDIA ACCESS or MULTIPLE ACCESS protocols, techniques, schemes, etc…) CS 515 © Ibrahim Korpeoglu

Wireless Media Access Control
Random Schemes (Less-Coordinated) Examples: MACA, MACAW, Aloha, MAC,… More suited for wireless networks that are designed to carry data: IEEE Wireless LANs Coordinated Schemes Examples: TDMA, FDMA, CDMA More suited for wireless networks that are designed to carry voice: GSM, AMPS, IS-95,… Polling based Schemes Examples: Bluetooth, BlueSky,… Access is coordinated by a central node Suitable for Systems that wants low-power, aims to carry voice and data at the same time. CS 515 © Ibrahim Korpeoglu

Duplexing It is sharing the media between two parties.
If the communication between two parties is one way, the it is called simplex communication. If the communication between two parties is two- way, then it is called duplex communication. Simplex communication is achieved by default by using a single wireless channel (frequency band) to transmit from sender to receiver. Duplex communication achieved by: Time Division (TDD) Frequency Division (FDD) Some other method like a random access method CS 515 © Ibrahim Korpeoglu

Duplexing Usually the two parties that want to communication in a duplex manner (both send and receive) are: A mobile station A base station Two famous methods for duplexing in cellular systems are: TDD: Time Division Duplex FDD: Frequency Division Duplex CS 515 © Ibrahim Korpeoglu

Duplexing - FDD F A duplex channel consists of two simplex channel with different carrier frequencies Forward band: carries traffic from base to mobile Reverse band: carries traffic from mobile to base M B R Base Station Mobile Station Reverse Channel Forward Channel fc,R fc,,F frequency Frequency separation Frequency separation should be carefully decided Frequency separation is constant CS 515 © Ibrahim Korpeoglu

Duplexing - TDD A single radio channel (carrier frequency) is shared in time in a deterministic manner. The time is slotted with fixed slot length (sec) Some slots are used for forward channel (traffic from base to mobile) Some slots are used for reverse channel (traffic from mobile to base) B M Base Station Mobile Station Slot number … channel F R F R F R F R …. Reverse Channel Forward Channel time Ti Ti+1 Time separation CS 515 © Ibrahim Korpeoglu

Duplexing – TDD versus FDD
FDD is used in radio systems that can allocate individual radio frequencies for each user. For example analog systems: AMPS In FDD channels are allocated by a base station. A channel for a mobile is allocated dynamically All channels that a base station will use are allocated usually statically. More suitable for wide-area cellular networks: GSM, AMPS all use FDD TDD Can only be used in digital wireless systems (digital modulation). Requires rigid timing and synchronization Mostly used in short-range and fixed wireless systems so that propagation delay between base and mobile do not change much with respect to location of the mobile. Such as cordless phones… CS 515 © Ibrahim Korpeoglu

Multiple Access - Coordinated
We will look now sharing the media by more than two users. Three major multiple access schemes Time Division Multiple Access (TDMA) Could be used in narrowband or wideband systems Frequency Division Multiple Access (FDMA) Usually used narrowband systems Code Division Multiple Access Used in wideband systems. CS 515 © Ibrahim Korpeoglu

Narrow- and Wideband Systems
Narrowband System The channel bandwidth (frequency band allocated for the channel is small) More precisely, the channel bandwidth is large compared to the coherence bandwidth of the channel (remember that coherence bandwidth is related with reciprocal of the delay spread of multipath channel) AMPS is a narrowband system (channel bandwidth is 30kHz in one-way) Wideband Systems The channel bandwidth is large More precisely, the channel bandwidth is much larger that the coherence bandwidth of the multipath channel. A large number of users can access the same channel (frequency band) at the same time. CS 515 © Ibrahim Korpeoglu

Narrowband Systems Wideband systems
Could be employing one of the following multiple access and duplexing schems FDMA/FDD TDMA/FDD TDMA/TDD Wideband systems Could be employing of the following multiple access and duplexing schemes CDMA/FDD CDMA/TDD CS 515 © Ibrahim Korpeoglu

Cellular Systems and MAC
Multiple Access Technique AMPS FDMA/FDD GSM TDMA/FDD USDC (IS-54 and IS-136) PDC CT2 Cordless Phone FDMA/TDD DECT Cordless Phone US IS-95 CDMA/FDD W-CDMA CDMA/TDD cdma2000 CS 515 © Ibrahim Korpeoglu

Frequency Division Multiple Access
B Individual radio channels are assigned to individual users Each user is allocated a frequency band (channel) During this time, no other user can share the channel Base station allocates channels to the users f1,F f2,F fN,F f1,R f2,R fN,R … M M M CS 515 © Ibrahim Korpeoglu

Features of FDMA An FDMA channel carries one phone circuit at a time
If channel allocated to a user is idle, then it is not used by someone else: waste of resource. Mobile and base can transmit and receive simultaneously Bandwidth of FDMA channels are relatively low. Symbol time is usually larger (low data rate) than the delay spread of the multipath channel (implies that inter-symbol interference is low) Lower complexity systems that TDMA systems. CS 515 © Ibrahim Korpeoglu

Capacity of FDMA Systems
Frequency spectrum allocated for FDMA system … Guard Band channel Guard Band Bt : Total spectrum allocation Bguard: Guard band allocated at the edge of the spectrum band Bc : Bandwidth of a channel AMPS has 12.MHz simplex spectrum band, 10Khz guard band, 30kHz channel bandwidth (simplex): Number of channels is 416. CS 515 © Ibrahim Korpeoglu

Time Division Multiple Access
The allocated radio spectrum for the system is divided into time slots In each slot a user can transmit or receive A user occupiess a cyclically repeating slots. A channel is logically defined as a particular time slot that repeats with some period. TDMA systems buffer the data, until its turn (time slot) comes to transmit. This is called buffer-and-burst method. Requires digital modulation CS 515 © Ibrahim Korpeoglu

TDMA Concept Downstream Traffic: Forward Channels: (from base to mobiles) 1 2 3 … N 1 2 3 …. N … Logical forward channel to a mobile Base station broadcasts to mobiles on each slot Upstream Traffic: Reverse Channels: (from mobile to base) 1 2 3 … N 1 2 3 …. N … Logical reverse channel from a mobile A mobile transmits to the base station in its allocated slot Upstream and downstream traffic uses of the two different carrier frequencies. CS 515 © Ibrahim Korpeoglu

TDMA Frames Multiple, fixed number of slots are put together into a frame. A frame repeats. In TDMA/TDD: half of the slots in the frame is used for forward channels, the other is used for reverse channels. In TDMA/FDD: a different carrier frequency is used for a reverse or forward Different frames travel in each carrier frequency in different directions (from mobile to base and vice versa). Each frame contains the time slots either for reverse channels or forward channel depending on the direction of the frame. CS 515 © Ibrahim Korpeoglu

General Frame and Time Slot Structure in TDMA Systems
One TDMA Frame Preamble Information Trail Bits Slot 1 Slot 2 Slot 3 … Slot N Guard Bits Sync Bits Control Bits Information CRC One TDMA Slot A Frame repeats in time CS 515 © Ibrahim Korpeoglu

A TDMA Frame Preamble contains address and synchronization info to identify base station and mobiles to each other Guard times are used to allow synchronization of the receivers between different slots and frames Different mobiles may have different propagation delays to a base station because of different distances. CS 515 © Ibrahim Korpeoglu

Efficiency of a Frame/TDMA-System
Each frame contains overhead bits and data bits. Efficiency of frame is defined as the percentage of data (information) bits to the total frame size in bits. bT: total number of bits in a frame Tf: frame duration (seconds) bOH: number of overhead bits Number of channels in a TDMA cell: m: maximum number of TDMA users supported in a radio channel CS 515 © Ibrahim Korpeoglu

TDMA TDMA Efficiency TDMA is usually combined with FDMA
GSM: 30% overhead DECT: 30% overhead IS-54: 20% overhead. TDMA is usually combined with FDMA Neighboring cells be allocated and using different carrier frequencies (FDMA). Inside a cell TDMA can be used. Cells may be re-using the same frequency if they are far from each-other. There may be more than one carrier frequency (radio channel) allocated and used inside each cell. Each carrier frequency (radio channel) may be using TDMA to further multiplex more user (i.e. having TDMA logical channels inside radio channels) For example: GSM uses multiple radio channels per cell site. Each radio channel has 200KHz bandwidth and has 8 time slots (8 logical channels). Hence GSM is using FHMA combined with TDMA. CS 515 © Ibrahim Korpeoglu

Contemporary TDMA Systems
GSM (Europa) IS-54 (USA) PDC (Japan) DECT (European Cordless) Bit Rate 270.8 Kbps 48.6 Kbps 42 Kbps 1.152 Mbps Bandwidth 200 KHz 30 KHz 25 KHz 1.728 MHz Time Slot 0.577 ms 6.7 ms 0.417 ms Upstream slots per frame 8 3 12 Duplexing FDD TDD Efficiency of Time Slots 73 % 80 % 67 % Modulation GMSK p/4 DQPSK Adaptive Equalized Mandatory Optional None CS 515 © Ibrahim Korpeoglu

Features of TDMA Enables the sharing of a single radio channel among N users Requires high data-rate per radio channel to support N users simultaneously. High data-rate on a radio channel with fixed bandwidth requires adaptive equalizers to be used in multipath environments (remember the RSM delay spread s parameter) Transmission occurs in bursts (not continues) Enables power saving by going to sleep modes in unrelated slots Discontinues transmission also enables mobile assisted handoff Requires synchronization of the receivers. Need guard bits, sync bits. large overhead per slot. Allocation of slots to mobile users should not be uniform. It may depend on the traffic requirement of mobiles. This brings extra flexibility and efficiency compared to FDMA systems. CS 515 © Ibrahim Korpeoglu

Capacity of TDMA Systems
Capacity can be expressed as System Capacity (the capacity of the overall system covering a region) Depends on: Range of cells Whether the system can support macro-cells, micro-cells or pico-cells. Cell Capacity Depends on the radio link performance between a base-sation and mobiles: The lowest C/I (carrier-to-interence) ratio the system can operate for example quality of transmission. This in turn depends on the speech coding technique, desired speech quality, etc. Data-rate over the channel which depends modulation efficiency (bits_per_second/Hz) and channel bandwidth. The frequency re-use factor CS 515 © Ibrahim Korpeoglu

System Capacity: Cluster: 7 cells constitute a cluster. C B
Cluster size = 7 A y B D A x G G E A z F Frequency reuse factor is 1/7: same frequency is used every 7 cell. A is one set of frequencies, B is an other, etc. A mobile in cell x receives carrier signal from base x and interferences from base stations at cells y and z. The carrier signal strength of all combined signal strength from interfering base stations is called C/I or S/I ratio. CS 515 © Ibrahim Korpeoglu

C/I affect on capacity C/I ratio affects the cluster size, hence the frequency reuse factor. Frequency_reuse_factor = 1 / cluster_size Cluster size can be 3, 7, 12, 13, … Cluster size affects the cell capacity (it affects the maximum number of frequencies that can be used in a cell) A low C/I requirement for appropriate quality enables smaller cluster sizes, hence larger frequency reuse factor, meaning that larger cell capacities CS 515 © Ibrahim Korpeoglu

Comparing Systems GSM Parameters AMPS Parameters
Channel banwidth: 200 KHz Required bandwidth per user = 200/8 = 25 Khz. Required minimum C/I: 9dB Frequency re-use factor: 1/3 (cluster size=3) AMPS Parameters Channel bandwidth = 30Khz Required C/I: 18 dB Frequency re-use factor: 1/7 (cluster size=7) Required bandwidth per user = 30kHz. CS 515 © Ibrahim Korpeoglu

Spread Spectrum Access
SSMA uses signals that have transmission bandwidth that is several orders of magnitued larger than minimum required RF bandwidth. Provides Immunity to multipath interference Robust multiple access. Two techniques Frequency Hopped Multiple Access (FHMA) Direct Sequence Multiple Access (DSMA) Also called Code Division Multiple Access – CDMA CS 515 © Ibrahim Korpeoglu

Frequency Hopping (FHMA)
Digital muliple access technique A wideband radio channel is used. Same wideband spectrum is used The carrier frequency of users are varied in a pseudo-random fashion. Each user is using a narrowband channel (spectrum) at a specific instance of time. The random change in frequency make the change of using the same narrowband channel very low. The sender receiver change frequency (calling hopping) using the same pseudo-random sequence, hence they are synchronized. Rate of hopping versus Symbol rate If hopping rate is greather: Called Fast Frequency Hopping One bit transmitted in multiple hops. If symbol rate is greater: Called Slow Frequency Hopping Multiple bits are transmitted in a hopping period GSM and Bluetooth are example systems CS 515 © Ibrahim Korpeoglu

Case Study - Bluetooth Uses Frequency Hopping in cell (piconet) over a 79 MHz wideband radio channel. Uses 79 narrowband channels (carrier frequencies) to hop through. Freq (f) = 2402+k MHz, k = 0,...,78 Channel spacing is 1 MHz (narrowband channel bandwidth) Wideband spectrum width = 79 MHz. Hopping Rate = 1600 Hops/Second Hopping sequence is determined by Bluetooth Hardware address and Clocks that are syncrozied between sender and receiver 79 MHZ 1 2 3 ..... 77 78 79-Hop System 1 MHZ A hop sequence could be: 7,1,78,67,0, 56,39, CS 515 © Ibrahim Korpeoglu

Case Study: Bluetooth – Piconet and FHSS
Each node is classified as master or slave. Master defines a piconet (a cell). Maximum 7 slaves can be connected to a master. Master coordinates access to the the media. All traffic has to go over master. Slaves can not talk to each-other directly. Picocell S Range = 10m Raw Data-rate: 1 Mbps/piconet Radio channel used by devices in a piconet is 79MHz channel, which Is frequency hopped: hopping though 789 channels. Hoprate = 1600 hops/sec FHSS M S S All slaves and the master hops according to the same hopping sequence. The hopping sequence is determined by the clock and BT_address of the master. CS 515 © Ibrahim Korpeoglu

Case Study: Bluetooth – Scatternet and FHSS
Piconet S S Piconet can be combined into scatternets. Red slave acts as a bridge between two piconets. M2 Piconet S FHSS S FHSS M1 Each piconet uses FHSS with different hopping sequences (masters are different). This prevents interference between piconets. S S CS 515 © Ibrahim Korpeoglu

Case Study: Bluetooth - Media access in a piconet
Inside a piconet, access to the frequency hopped radio channel is coordinated using time division multiple access: TDMA/TDD. Slot duration = 1/1600 sec = 625ms Piconet S1 FHSS M In an even slot, master transmits to a slave. In an odd slot, the slave that is addressed in the previous master-to-slave slot transmits. S3 S2 ….. M-S1 S1-M M-S2 S2-M M-S3 S3-M M-S1 S1-M …… slot time=625ms CS 515 © Ibrahim Korpeoglu

Code Division Multiple Access (CDMA)
In CDMA, the narrowband message signal is multiplied by a very large bandwidth signal called spreading signal (code) before modulation and transmission over the air. This is called spreading. CDMA is also called DSSS (Direct Sequence Spread Spectrum). DSSS is a more general term. Message consists of symbols Has symbol period and hence, symbol rate Spreading signal (code) consists of chips Has Chip period and and hence, chip rate Spreading signal use a pseudo-noise (PN) sequence (a pseudo-random sequence) PN sequence is called a codeword Each user has its own cordword Codewords are orthogonal. (low autocorrelation) Chip rate is oder of magnitude larger than the symbol rate. The receiver correlator distinguishes the senders signal by examining the wideband signal with the same time-synchronized spreading code The sent signal is recovered by despreading process at the receiver. CS 515 © Ibrahim Korpeoglu

CDMA Advantages Low power spectral density.
Signal is spread over a larger frequency band Other systems suffer less from the transmitter Interference limited operation All frequency spectrum is used Privacy The codeword is known only between the sender and receiver. Hence other users can not decode the messages that are in transit Reduction of multipath affects by using a larger spectrum Random access possible Users can start their transmission at any time Cell capacity is not concerete fixed like in TDMA or FDMA systems. Has soft capacity Higher capacity than TDMA and FDMA No frequency management No equalizers needed No guard time needed Enables soft handoff CS 515 © Ibrahim Korpeoglu

CDMA Principle Represent bit 1 with +1 Represent bit 0 with -1
One bit period (symbol period) 1 1 Data PN-Code (codeword) Coded Signal Input to the modulator (phase modulation) Chip period CS 515 © Ibrahim Korpeoglu

Processing Gain Main parameter of CDMA is the processing gain that is defined as: Gp: processing gain Bspread: PN code rate Bchip: Chip rate R: Data rate IS-95 System (Narrowband CDMA) has a gain of 64. Other systems have gain between 10 and 100. 1.228 Mhz chipping rate 1.25 MHz spread bandwidth CS 515 © Ibrahim Korpeoglu

Near Far Problem and Power Control
At a receiver, the signals may come from various (multiple sources. The strongest signal usually captures the modulator. The other signals are considered as noise Each source may have different distances to the base station In CDMA, we want a base station to receive CDMA coded signals from various mobile users at the same time. Therefore the receiver power at the base station for all mobile users should be close to eacother. This requires power control at the mobiles. Power Control: Base station monitors the RSSI values from different mobiles and then sends power change commands to the mobiles over a forward channel. The mobiles then adjust their transmit power. B pr(M) M M M M CS 515 © Ibrahim Korpeoglu

DSSS Transmitter Message Baseband sss(t) + BPF m(t) Transmitted Signal
p(t) PN Code Generator Oscillator fc Chip Clock CS 515 © Ibrahim Korpeoglu

DSSS Receiver Phase Shift Keying IF Wideband Demodulator Filter
Received Data Received DSSS Signal at IF PN Code Generator Synchronization System CS 515 © Ibrahim Korpeoglu

Spectra of Received Signal
Spectral Density Spectral Density Interference Signal Interference Signal Frequency Frequency Output of Correlator after dispreading, Input to Demodulator Output of Wideband filter CS 515 © Ibrahim Korpeoglu

CDMA Example (*) R Receiver (a base station) Data=1011… Data=0010… A B
Transmitter (a mobile) Transmitter Codeword=101010 Codeword=010011 Data transmitted from A and B is multiplexed using CDMA and codewords. The Receiver de-multiplexes the data using dispreading. (*) This example is adapted from the CDMA example of Prof. Randy Katz at UC-Berkeley. CS 515 © Ibrahim Korpeoglu

CDMA Example – transmission from two sources
A Data A Codeword A Signal B Data B Codeword B Signal Transmitted A+B Signal CS 515 © Ibrahim Korpeoglu

CDMA Example – recovering signal A at the receiver
A+B Signal received A Codeword at receiver Integrator Output Comparator Output Take the inverse of this to obtain A CS 515 © Ibrahim Korpeoglu

CDMA Example – recovering signal B at the receiver
A+B Signal received B Codeword at receiver Integrator Output Comparator Output Take the inverse of this to obtain B CS 515 © Ibrahim Korpeoglu

CDMA Example – using wrong codeword at the receiver
A+B Signal received Wrong Codeword Used at receiver Integrator Output Comparator Output X Noise Wrong codeword will not be able to decode the original data! CS 515 © Ibrahim Korpeoglu

Hybrid Spread Spectrum Techniques
FDMA/CDMA Available wideband spectrum is frequency divided into number narrowband radio channels. CDMA is employed inside each channel. DS/FHMA The signals are spread using spreading codes (direct sequence signals are obtained), but these signal are not transmitted over a constant carrier frequency; they are transmitted over a frequency hopping carrier frequency. CS 515 © Ibrahim Korpeoglu

Hybrid Spread Spectrum Techniques
Time Division CDMA (TCDMA) Each cell is using a different spreading code (CDMA employed between cells) that is conveyed to the mobiles in its range. Inside each cell (inside a CDMA channel), TDMA is employed to multiplex multiple users. Time Division Frequency Hopping At each time slot, the user is hopped to a new frequency according to a pseudo-random hopping sequence. Employed in severe co-interference and multi-path environments. Bluetooth and GSM are using this technique. CS 515 © Ibrahim Korpeoglu

Random Access Packet Radio Protocols CSMA Protocols
Multihop radio network that carries packets Not circuit oriented like GSM, CDMA, etc. Example Protocols Pure Aloha Slotted Aloha CSMA Protocols 1-persistent CSMA non-persistent CSMA p-persistent CSMA CSMA/CD Reservation Protocols Reservation Aloha PRMA Others MACA, MACAW IEEE MAC CS 515 © Ibrahim Korpeoglu

Pure Aloha Algorithm: A mobile station transmits immediately whenever is has data. It then waits for ACK or NACK. If ACK is not received, it waits a random amount of time and retransmits. Ignoring the propagation delay between mobiles and base station: B The time difference between the time a mobile send the first bit of packet and the time the base station receives the last bit of the packet is given by 2T. T = C/P T: packet time. C: channel data rate (bps) P: packet length (bits) Ack/Nack Data M1 M3 M2 During this 2T period of time, the packet may collide with someone elses packet. CS 515 © Ibrahim Korpeoglu

CSMA: Carrier Sense Multiple Access
Aloha does not listen to the carrier before transmission. CSMA listen to the carrier before transmission and transmits if channel is idle. Detection delay and propagation delay are two important parameters for CSMA Detection delay: time required to sense the carrier and decide if it is idle or busy Propagation delay: distance/speed_of_ligth. The time required for bit to travel from transmitter to the receiver. CS 515 © Ibrahim Korpeoglu

CSMA Variations 1-persistent CSMA: Non-persistent CSMA:
A station waits until a channel is idle. When it detects that the channel is idle, it immediately starts transmission Non-persistent CSMA: When a station receives a negative acknowledgement, it waits a random amount of time before retransmission of the packet altough the carrier is idle. P-persistent CSMA P-persistent CSMA is applied to slotted channels. When a station detects that a channel is idle, it starts transmission with probability p in the first available timeslot. CSMA/CD Same with CSMA, however a station also listen to the carrier while transmitting to see if the transmission collides with someone else transmission. Can be used in listen-while-talk capable channels (full duplex) In single radio channels, the transmission need to be interrupted in order to sense the channel. CS 515 © Ibrahim Korpeoglu

MACA – Medium Access with Collision Avoidance
CSMA protocols sense the carrier, but sensing the carrier does not always releases true information about the status of the wireless channel There are two problems that are unique to wireless channels (different than wireline channels), that makes CSMA useless in some cases. These problems are: Hidden terminal problem Exposed terminal problem. CS 515 © Ibrahim Korpeoglu

MACA – Hidden Terminal Problem
A’s cell C’s cell A B C Hidden terminal A is transmitting to B. C is sensing the carrier and detects that it is idle (It can not hear A’s transmission). C also transmits and collision occurs at B. A is hidden from C. CS 515 © Ibrahim Korpeoglu

MACA – Exposed Terminal Problem
B’s cell C’s cell A B C D Exposed terminal B is transmitting to A. C is hearing this transmission. C now wants to transmit to D. It senses the existence of carrier signal and defers transmission to D. However, C can actually start transmitting to D while B is transmitting to A, Since A is out of range of C and C’s signals can not be heard at A. C is exposed to B’s transmission. CS 515 © Ibrahim Korpeoglu

Ali, lets talk! I am available.
MACA Solution Concept Ali, lets talk! I am available. Can Can, I want to talk to you! Can, I want to talk to you! Biltepe Mountain Ali Veli CS 515 © Ibrahim Korpeoglu

MACA Protocol When a station wants to transmit data
It sends an RTS (Ready-to-Send) packet to the intended receiver The RTS packet contains the length of the data that needs to be transmitted Any station other than the intended recipient hearing RTS defers transmission for a time duration equal to the end of the corresponding CTS reception The receiver sends back CTS (Clear-to-Send) packet back to sender if it is available to receive. The CTS packet contains the length of the data that original sender wants to transmit Any station other than the original RTS sender, hearing CTS defers transmission until the data is sent. The original sender upon reception of the CTS, starts transmitting. CS 515 © Ibrahim Korpeoglu

MACA Solution for Hidden Terminal Problem
A is transmitting to B. A’s cell C’s cell RTS(n) RTS(n) CTS(n) X A B C X defers transmission until expected CTS reception time by RTS sender. CTS(n) C defers transmission for duration of n bytes of data transmission. Node A is no longer hidden from C effectively. Data(n) Waiting time of node X is much smaller than waiting time of node C. CS 515 © Ibrahim Korpeoglu

MACA Solution for Exposed Terminal Problem
B is transmitting to A B’s cell C’s cell RTS(n) RTS(n) A B C D RTS(m) CTS(n) CTS(m) Data(n) Data(m) C defers transmission upon hearing B’s RTS until B could get CTS from A. After that C can start transmission to D. For that it first sends an RTS. C is not longer exposed to the data transmission of B. CS 515 © Ibrahim Korpeoglu

Case Study: IEEE b MAC IEEE b: High Data-rate Wireless LAN standard. Operates in MHz ISM RF Band. 83 MHz spectrum width Max data-rate: 11Mbps simplex. Spectrum Usage: FHSS or DHSS Modulation Technique: CCK with QPSK For 11Mbps: Symbol rate = 1,375 MSps Number of symbol states = 8 One symbol can encode 3 bits of information. Range: around 100m. CS 515 © Ibrahim Korpeoglu

802.11b Infrastructure Mode Works in Two Operational Modes
Ad-Hoc Mode Infrastructure Mode Access Point Access Point Wireless Link Wireless Link Wireless Link Mobile Station Extended Service Set (ESS) Basic Service Set (BSS) All traffic has to go through access points Access point provides connectivity to the wired backbone CS 515 © Ibrahim Korpeoglu

802.11b Ad-Hoc Mode Independent Basic Service Set (IBSS)
Mobile Stations can talk directly with each-other. All stations in an IBSS need to be in the range of each-other. CS 515 © Ibrahim Korpeoglu

802.11b MAC Sublayer Support two different MAC modes depending on the operational mode of the Wireless LAN 1) DCF: Distributed Coordination Function Based on CSMA/CA Carrier Sensing: Physical and Virtual. 2) PCF: Point Coordination Function Connection oriented Contention free service Polling based CS 515 © Ibrahim Korpeoglu

802.11b PHY Layer 802.11b Data Rate Specifications
Can support data rates at: 1,2,5.5,11 Mbps 802.11b Data Rate Specifications Data rate Code Length Modulation Symbol Rate Bits/Symbol 1 Mbps 11 (Barker Sequence) BPSK 1 MSps 1 2 Mbps QPSK 2 5.5 Mbps 4 (CCK) 1.375 MSps 4 11 Mbps 8 (CCK) 8 CS 515 © Ibrahim Korpeoglu

FHSS 2.4 GHz band is divided into 75 one-MHz subchannels. The sender and receiver hops through this 75 channels in a synchronized manner using a hopping pattern. Can not support more than 2 Mbps data-rate. CS 515 © Ibrahim Korpeoglu

DSSS Divides the 2.4 GHz band into 14 twenty-two MHz channels
Adjacent channels can overlap partially. 3 of 14 channels are completely non-overlapping Data is sent over one 22 MHz channel without hoppling using DSSS technique (chipping and code words are used like CDMA) Each access point uses a different 22 MHz channel if possible. All mobiles in the coverage of the access point uses the channel that is used by the access point b MAC is used to coordinate the access to the shared 22 MHz channel. Original systems use 11 bit chipping (code words of length 11). Later b systems use 8 bit chipping (code words of length 8 bits). Defines 64 different codewords from a space of 256. CS 515 © Ibrahim Korpeoglu

Spectrum Allocated for 802.11b
DSSS Channels 25 MHz 25 MHz 2.412 GHz 2.437 GHz 2.462 GHz Channel 1 22 MHz Channel 6 22 MHz Channel 11 22 MHz 2.400 GHz 2.484 GHz Spectrum Allocated for b Channel 1, 6, and 11 are non-overlapping. CS 515 © Ibrahim Korpeoglu

Channel Assignment and Registration.
In multi-access environment, the operator should try to allocate non-overlapping channels to the physically adjacent channels. If adjacent access points use overlapping channels, then interference can be high. A mobile station periodically tunes to all channels and evaluates the signal strength received over each channel Depending on the signal strength received over the channels, a mobile selects an access point and registers with that provided that the access points accepts the mobile. This is also called association. Re-association with a new access point occurs when the mobile moves away from the current access point. When the signal conditions changes between the mobile and current access point. When there are a lot of users associated with the current access point. CS 515 © Ibrahim Korpeoglu

Re-association at the PHY layer.
Access Point (AP) A Access Point (AP) B Signal from A Signal from B Associated with Access Point B Associated with Access Point A Mobile tunes to the channel of AP B when it moves into its range. CS 515 © Ibrahim Korpeoglu

An example 3-cell Reuse scheme for WLAN deployment
1 11 11 6 6 1 1 1 11 11 6 6 An access point is located in the center of each hexagon. CS 515 © Ibrahim Korpeoglu

MPDU (Variable Length)
802.11b PHY Layer MAC PLCP: Physical Layer Convergence Protocol PMD: Physical Medium Dependent Sublayer PLCP PHY Layer PMD Sublayer PLCP Frame Format SYNC (128) SFD (16) Signal (8) Service (8) Length (16) CRC (16) MPDU (Variable Length) SYNC: Synchronization field SFD: Start frame deliminer Signal: Indicated how fast the data will be transmitted Service: Reserved Length: MAC Protocol Data Unit (MPDU) length CRC: used for error detecting on the frame CS 515 © Ibrahim Korpeoglu

802.11b MAC Sublayer Supports both infrastructure and ad-hoc modes of operation. CRC is added to each MAC frame Packet fragmentation is supported to chop large higher layer (IP) packets into small pieces. Has advantages: Probability a packet gets corrupted increases with the packet size. In case of corruption, only a small fragment needs to be re-transmitted. CS 515 © Ibrahim Korpeoglu

Inter-frame Space (IFS
4 types of Inter-frame spaces: Short IFS (SIFS): period between completion of packet transmission and start of ACK frame Point Coordination IFS (PIFS): SIFS plus a slot time. Distributed IFS (DIFS): PIFS plus a slot time. Extended IFS (EIFS): longer IFS used by a station that has received a packet that it could no longer understand. Needed to prevent collisions. CS 515 © Ibrahim Korpeoglu

MAC Protocol 802.11b uses CSMA/CA (Carrier Sense Multiple Access/Collision Avoidance) MAC protocol. CSMA/CA is the protocol to implement the distributed coordination function (DCF) of the MAC sub-layer. RTS/CTS is used to avoid collisions. Use of RTS/CTS can be enabled or disabled depending on the traffic load (probability of collisions). CS 515 © Ibrahim Korpeoglu

CSMA – Transmission of MPDU (Data) without use of RTS/CTS
DIFS Source Data SIFS Destination ACK Contention Window (Slot Times) DIFS Others Data Defer Access Backoff after Defer A station backoffs a random number of slot times. CS 515 © Ibrahim Korpeoglu

CSMA/CA – Transmission of MPDU (Data) using RTS/CTS
DIFS Source RTS DATA SIFS SIFS SIFS Destination CTS ACK Others DIFS Defer Access for NAV(RTS) Defer Access for NAV(CTS) Backoff after Defer Defer Access for NAV(Data) CS 515 © Ibrahim Korpeoglu

CSMA/CA Collision Avoidance
RTS/CTS is used to reserve channel for the duration of the packet transmission. This prevents hidden and exposed terminal problems ACK is required to understand if the packet is correctly received (without any collisions ) at the receiver. Ethernet does not require ACK to be sent, since the transmitter can detect the collision on the channel (cable) without requiring an explicit feedback from the receiver. A wireless transmitter can not detect collision, because: 1) Transmit power is much larger than the received power: received signal is regarded as noise (not collision). 2) There could be a hidden terminal Access Point Mobile RTS CTS DATA ACK CS 515 © Ibrahim Korpeoglu

802.11b Frame Format IEEE 802.11b MAC Frame Format FC (2 bytes) ID (2)
Add1 (6) Add2 (6) Add3 (6) SC (2) Add4 (6) Data ( bytes) CRC (4) Frame Control Format (2 bytes) Protocol (2 bits) Type (2) Subtype (4) To DS (1) From DS (1) More Frag (1) Retry (1) Pw Mgt (1) More Data (1) WEP (1) Order (1) Protocol Version: version of standard Type: Management. Control, Data frame Subtype: RTS, CTS, ACK frame To DS: 1 if frame is sent to Distribution System (DS) From DS: 1 if frame is received from Distribution System More fragment: 1 if there are more fragments belonging to the same frame following the current frame. Retry: indicates that is fragment is retransmission of previously transmitted fragment. Power Management: the type of power management mode that the station will be after the transmission of the frame. More Data: indicates that there are more frames buffered at the sender for this station. WEP: indicates that frame body is encrypted according to WEP. Order: indicates that the frame is sent using the strictly-ordered service class. Frame Control (FC): protocol version and frame type Duration/ID (ID): power-save poll message frame type and for NAV calculation Address Fields: contains up-to 4 MAC addresses Sequence Control: fragmentation and sequence number. Data: higher layer data that is maximum bytes. CRC: 32 bit cyclic redundancy check for detecting error on the frame. CS 515 © Ibrahim Korpeoglu

Mobility What happens when a station moves between access points
Re-association function of the PHY layer associates a mobile with a new access point. Some vendor specific, layer-2 (datalink layer) solutions solves the mobility at layer. Solutions like Mobile IP needed to provide seamless mobility to higher layers (transport and application layers). DHCP is also a method but not as convenient as Mobile IP. We will see in the forthcoming classes how Mobile IP works. CS 515 © Ibrahim Korpeoglu

Mobility Support in Internet and Mobile IP
CS Mobile and Wireless Networking İbrahim Körpeoğlu Computer Engineering Department Bilkent University Bilkent / ANKARA İbrahim Körpeoğlu

Problem We have seen that mobile users can change point of attachment
In a WLAN, a mobile changes access point. In a cellular network, a mobile changes base station. A mobile user can work at office and at home at different in a day: mobile changes Ethernet subnets. A mobile PDA user may connect to its ISP using a modem and PPP protocol from different telephone lines (telephone jacks) at different places: home, work, a foreign location. We want out applications to be not disturbed from mobility We want to continue to talk with our cell-phone when we change base-stations We want to continue to run telnet when we change access points in a Wireless LAN. .... CS 515 © İbrahim Körpeoğlu

Two kinds of mobility 1) Mobility is totally transparent to applications This is called seamless mobility 2) Mobility is not transparent to applications when we move, but we can still access the network at a new place. This is called portability Some protocols support either one of them Mobile IP can support seamless mobility DHCP can support portability CS 515 © İbrahim Körpeoğlu

Mobility Solutions Mobile Cellular Telephone Networks and Mobile Internet has different protocols and solutions to support mobile users. Mobile Cellular Telephone Networks Solution GSM has its own registration, handoff, mobility management procedures Mobile Internet Solution Mobile IP has been developed to support IP based hosts and mobile users. We will look to Mobile IP first. CS 515 © İbrahim Körpeoğlu

Mobile IP Mobile IP is a layer-3 (network layer) mobility solution to support mobile users (laptops, etc) in the Internet in a seamless manner. By use of Mobile IP, all TCP/IP applications (applications that use sockets) are unaware of the fact that the users are moving and changing their piont of attachment to the Internet. Only IP protocol and lower layers are aware of mobility Higher protocol layers (TCP, UDP, RTP, etc) and applications are not aware of mobility. CS 515 © İbrahim Körpeoğlu

Mobile IP We concetrate in how mobility support is done at the network layer. We will not be concerned about how mobile stations change physical point of attachments at the Physical layer. This depends on the Physical Media: We have seen how this is achieved in Wireless LAN (802.11b) protocol: re-association with a new access point when the signals get weaker. In Ethernet, we just need to plug out the cable from an old attachment point (jack or HUB) to a new point (a new hub) to change Physical attachment. Other Physical layer may have other procedures to change the point of attachment. Mobile IP is a solution that is independent of the physical and data-link layers: It can work for Ethernet, Token Ring, Wireless LANs, PPP over serial cables or phone lines,, etc. CS 515 © İbrahim Körpeoğlu

Mobile IP Network Applications and Protocols Telnet, FTP, HTTP, etc.
Mobility is seamless to these. TCP UDP TCP/IP Protocol STACK Mobile IP Ethernet Token Ring PPP WLAN 802.11b Bluetooth .... The link layer can be anything. CS 515 © İbrahim Körpeoğlu

Why wee need mobile IP Current Internet architecture and protocols (without mobile IP support) do not support seamless mobility for mobile users Internet is designed assuming hosts (computer) are static and do not change location frequently. When you move to a new location with your laptop and connect it to a Ethenet cable at the new location, you have re-confıgure your laptop. Obtain new IP address, Learn the subnet mask. Learn the deafult router IP address Learn the local DNS servers IP addresses When you re-confıgure your laptop with this information, most of the time you have re-start your laptop. Whether you re-start or not your laptop, previously running network applications will stop working properly when you change the IP address of your laptop. CS 515 © İbrahim Körpeoğlu

Why wee need mobile IP Initially we had desktops, workstations, main-frames and super-computer all of whic are static and heavy enough so that you can not carry them with you!. Initial design of Internet was for these computers. Now, we have Laptop and handheld computers which you carry to new places when you travel Palmtop and Pocket PC computers which you carry in your pocket even if you go to a movie. An these are powerful enough to run a lof interesting network applications like web browsers, etc. Hence you still need Internet access for these highly mobile computers and devices That is why wee need mobility support to be added to the Internet. Mobile IP has been designed for this purpose! CS 515 © İbrahim Körpeoğlu

Problems with Internet for Mobility
In Internet, IP addreses are used for two purposes Identification of hosts Both an IP address or domain name address (FQDN) can be used to identify a host. DNS servers does the mapping between IP addresses and domain names Usually there is one to one mapping. Network protocol in TCP/IP stack usually use IP addresses to identify the end-point Applications may use the domain names so that they are more user friendly to the humans. Locating mobile hosts: for Routing IP addresses are structured and correspond to well-specific locations in Internet. They are used for detemining the routes that packets will follow from a source machine to a destination machine. For static hosts, we can use its IP address for very long times, since the location dependent IP address does not have to be changed, since a static host do not change location. CS 515 © İbrahim Körpeoğlu

Problems with Internet for Mobility
When mobile hosts come into picture in Internet: We need a location-independent identifier for the mobile hosts so that any user who wants to contact to the mobile host should be able to use this identifier to send information to the mobile host without getting bothered with the current location of the mobile. We also need a new location-dependent IP address (all IP addresses are location-dependent) for a mobile host when it moves to a new location in order to route the packets destined for the mobile to the new location so that the mobile can receive them at the new location. Hence, a single IP address for the a mobile host can not serve both purposes (identity and location/routing) at the same time. CS 515 © İbrahim Körpeoğlu

Mobile IP Approach Use two IP addresses per mobile host
One permanent IP address (also called home-address) Used for Identification An other IP address that is changing depending on the current location the mobile host (called care-of-address) Used for Routing The binding (association) between these two IP addresses are kept at a well-known location, called home agent. CS 515 © İbrahim Körpeoğlu

Why DHCP is not enough DHCP: Dynamic Host Configuration Protocol
An Internet Protocol that allows host that does not have an IP address to obtain an IP address and other configuration information when it connects to a network at a new location. Network to be connected can be for example an Ethernet link Network to be connected should support DHCP protocol The mobile host should support DHCP protocol The configuration info that can be obtained via DHCP at the new location includes: A registered IP address Subnet mask of the network Local DNS server IP addresses (primary and secondary IP addresses), ... CS 515 © İbrahim Körpeoğlu

Example Assume we have DHCP support in CS department, Math department and dormitories. Assume you have a laptop that has DHCP support installed. You don’t need to obtain an IP address from BCC in this case. You don’t need to bother with network configuration of your laptop. You will just plug-in your laptop to an Ethernet jack at CS department, at Math department, or at your dormitory and you will be online instantly and easily. You can move around between CS and Math departments and your dormitory together with your laptop and get connected to the network. Disadvantage You have to reboot you computer whenever you connect it to a new network (ethernet jack at a new location). All applcations have to be restarted. You laptop obtains a new IP address at the new location from DHCP server. You can connect to outside world with this new IP address. However, Your friends wil not able to contact to you. Mobility is not seamless. CS 515 © İbrahim Körpeoğlu

DHCP does not provide seamless mobility
Since you obtain a new IP address a every new location, applications has to be restarted Restart is not problem for web page access Restart is problem for telnet and ftp sessions and some other network and TCP applications. Other people can not connect to you when you move to a new location unless they learn your new IP address You have to call them and let your IP address at every move!!! DNS servers are not dynamic enough currently to update the binding between your machine’s domain name (host name) and its IP address. This binding will be stale when you move to a new location. Your friend who wants to contact to you and uses your machine’s host name, will have the old IP address returned from the DNS server. Hence the packets (messages) he will sent will be routed to your old IP address. CS 515 © İbrahim Körpeoğlu

Mobile IP Protocol Overview
2 3 Home Agent Foreign Agent 4 Mobile Node Internet 1 5 Correspondent Host CS 515 © İbrahim Körpeoğlu

Mobile IP Functions Agent Discovery Registration Tunneling
When a mobile node moves into a new subnetwork (or network), lt has to discover the foreign agent in that network For this, mobile agents (home and foreign) advertise their presence periodically using ICMP messages. Registration When a mobive moves to a new network and obtains a new care-of-address there, ıt has to register that address with the home agent (binding), so that home agent knows where to forward the packets aimed for mobile. Registration should be secure Tunneling When packet aimed for mobile are intercepted by home agent, they are forward to the current location (care-of-address) of the mobile using a mechanism called Tunneling There are various forms of tunneling: IP-IP, Minimum Encapsulation, GRE, etc. CS 515 © İbrahim Körpeoğlu

Example A correspondent host C wants to send an IP packet to a mobile host M. It generated the IP packet so that the IP packet has destination address equal to mobile’s home address The IP packet is send to the mobile’s home address Routers forward the packet using normal Internet routing mechanisms to the home network of the mobile. Assume mobile is away from home network and currenty is located in a foreign network. Hence mobile will not be able to receive (capture) the packet that is sent to the mobile’s home network. A home agent located in the mobile’s home network will intercept the packet aimed for mobile This interception is done with the help of proxy ARPing. Home agent will know the whereabouts of the mobile, if the mobile has registered with the home agent previously. CS 515 © İbrahim Körpeoğlu

Example – continued. Home agent will encapsulate the IP packet using IP-IP encapsulation (tunneling) method and will send the encapsulated IP packet to the new location (care-of-address) of the mobile. The new location is the foreign network that the mobile currently resides in. The encapsulated IP packet will be transported to the care-of-address of the mobile using normal Internet routing mechanisms. Care-of-address can be the IP address of a foreign agent or the new IP address of the mobile at thew new location obtained via methods like DHCP, etc. In this case the foreign agent could be co-located at the mobile host. The holder of the care-of-address (a foreign agent) will receive the encapsulated IP datagram, wil strip off the outer header (decapsulate) and will forward the original IP packet to the mobile host. The mobile host will receive the packet as it is coming from a correspondent host directly without going through the home agent (if foreign agent functionality is not co-located at the mobile host). CS 515 © İbrahim Körpeoğlu

IP-C: IP Address of Correspondent Host
Mobile Host C IP-C: IP Address of Correspondent Host IP-M: IP Adress of Mobile Host (home address of mobile) IP-H: IP Address of Home Agent (care-of-address of mobile) IP-F: IP Address of Foreign Agent. Dst Src IP Payload IP-M IP-C …. Home Agent H Tunnel Dst Src Dst Src IP Payload IP-M IP-C …. IP-F IP-H …. Foreign Agent F Inner IP Header Outer IP Header Other Fields INTERNET Dst Src An IP Header Fields IP Payload IP-M IP-C …. Ver HL TOS Total Length IP Header Identification Flags Fragm. Offset TTL Protocol Header Checksum Correspondent Host C Src Address Dest Address Packet Transport from a Correspondent Host to a Mobile CS 515 © İbrahim Körpeoğlu

IP-C: IP Address of Correspondent Host IP-M: IP Adress of Mobile Host
IP-H: IP Address of Home Agent IP-F: IP Address of Foreign Agent. Src Dst IP Payload IP-M IP-C …. Home Agent H Foreign Agent F INTERNET Src Dst IP Payload IP-M IP-C …. Correspondent Host C Packet Transport from a Mobile to a Correspondent Host CS 515 © İbrahim Körpeoğlu

Mobile Agent Discovery
How a mobile node discovers the home and foreign agents when it travels? Agents periodically broadcast their presence (advertisement) on a link ( a wireless link – , or a wired link – ethernet) These broadcasts are agent advertisement messages. A mobile node receiving the advertisement understand from the IP addresses included in the advertisement: Whether it is in the home network or not? Whether it has moved to new location or not. This understanding is at the IP level (A mobile already knows that it has moved at the physical link level if has moved). CS 515 © İbrahim Körpeoğlu

Mobile Agent Discovery
An agent advertisement message is an ICMP router advertisement message with special extension. The special extension is called Mobility Agent Extension. CS 515 © İbrahim Körpeoğlu

Agent Advertisement Message
Ver HL TOS Total Length Identification Flags Fragm. Offset TCP/IP Protocol Stack in a Host TTL Protocol Header Checksum IP Header Src Address Applications Dest Address Type Code Checksum ICMP Router Advertisement Message TCP UDP NAddr=0 Addr Size Lifetime Type Length Sequence Number ICMP IGMP Mobility Agent Extension IP Lifetime Flags Reserved Zero or more care-of-addresses ………. ARP RARP Link Layer FLAGS R: Registration requires (with the foreign agent) B: Foreign agent is busy H: The agent is home agent. F: The agent is foreign agent M: Minimum encapsulation G: GRE encapsulation V: Van Jacobson Header Compression CS 515 © İbrahim Körpeoğlu

Registration After a mobile detects at the IP (ICMP) layer that it has moved to a new location, it starts registration procedure with the home agent. The aim of the registration is to let the home agent know mobile’s current care-of-address. Mobile obtains this care-of-address ether from the foreign agent or from a server like DHCP server. Registration procedure consists of sending a Registration Request Message from mobile to home agent and a Registration Reply Message from home agent to mobile Registration messages has to go through Foreign agent. Foreign Agent just forwards these registration messages back and forth Foreign agent is a passive entity in registration. . Registration messages sent over UDP to port number 434. CS 515 © İbrahim Körpeoğlu

Registration Request 0 8 16 31 HA REQ FA Type Flags Lifetime REQ
HA REQ FA Type Flags Lifetime REQ Home address Type: Type of the Mobile IP Message: 1 – Registration Request. Lifetime: Number of seconds registration is valid. Home address: The home IP address of the mobile Home agent: The IP address of the home agent. Care-of-address: The current IP address of the mobile – this is then end of the tunnel. Identification: Used for replay protection. Extensions: Security extensions can be added to protect from malicious people. Flags: S: Simultaneous binding. B: Broadcast – Home agent will tunnel broadcast datagrams to the mobile D: Mobile node is using a collocated care-of-address – that means there is no foreign agent and mobile node will decapsulate the packets itself. M: Mobile node requests the home agent to encapsulate the packets using Minimal Encapsulation G: Mobile node requests the home agent to encapsulate the packets using GRE Encapsulation M Home agent Care-of--address Identification Extensions ….. Registration Request Format IP Header UDP Header Mobile IP Message Extensions CS 515 © İbrahim Körpeoğlu

Registration Reply RPL HA FA RPL 0 8 16 31 M Type Code Lifetime
M Type Code Lifetime Home address Home agent Type: 3 – Registration Reply Code: Indicates the result of registration Some code values: 0 registration accepted 66 insufficient resources at foreign agent 70 poorly formed request 130 insufficient resources at home agent 131 mobile node failed authentication Lifetime: The granted life time by home agent for registration Identification Extensions ….. Registration Reply Format CS 515 © İbrahim Körpeoğlu

Care-of-Address Types
Normal Care-of-address The care-of-address that mobile obtains at a new location is the IP address of a foreign agent serving at that new location. Registration and communication has to go through foreign agent Collocated care-of-address There is no separate foreign agent present at the new location Mobile obtains an IP at the new location through some standard mechanisms like DHCP. This IP address is called collocated IP address. The foreign agent functionality is executed at the mobile node itself. The mobile node decapsulates the tunneled packets coming from home agent. Registration and communication is done directly between mobile and home agent. CS 515 © İbrahim Körpeoğlu

Securing the registration procedure
Security problem Fraudulent registrations should be detected. A bad person can send registration packets to home agent as if the packets are coming from a legitimate mobile user. In this way, the bad user can redirect the traffic destined to mobile node to itself and obtain the packets. Hence we need authentication There are three authentication extensions defined for Mobile IP The mobile-home authentication extension The mobile-foreign authentication extension The foreign-home authentication extension. CS 515 © İbrahim Körpeoğlu

Securing the registration procedure
Type: 32 – Mobile-Home authentication extension 33 – Mobile-Foreign authentication extension – Foreign-Home authentication extension SPI: Security Parameter Index. Defines the security context (algorithm, mode, key) to computer the authenticator. Authenticator: variable length. Type Length SPI SPI….continued Authenticator Authenticator….. Mobile IP Authentication Extension Added to the Registration Request Message Default Authentication Algorithm: Keyed-MD5 in prefix-suffix mode 128 bit authenticator: message digest of the registration message. Computer over: shared secret key, spi index, protected fields of registration message, shared secret again. CS 515 © İbrahim Körpeoğlu

Routing and Tunneling When a correspondent host sends an IP packet to a mobile (to its home address), packet is routed first to home agent of mobile through normal routing. Home agent intercepts the packet and encapsulates it and tunnels it to the care-of-address (tunnel exit point) of the mobile. The encapsulated packet is delivered to the care-of-address using normal routing. There are various encapsulation methods: IP-IP Encapsulation Minimal Encapsulation GRE (Generic Routing Encapsulation) Encapsulation. C Tunnel HA FA M Encapsulated IP Packet CS 515 © İbrahim Körpeoğlu

IP-IP Encapsulation at Home Agent
Ver HL TOS Total Length Home agent encapsulated the IP Packet inside an other IP header and Sends it to the care-of-address of mobile Identification Flags Fragm. Offset Outer Header TTL Protocol=4 Header Checksum Src Address = Home agent addres Dest Address = Care-of-Address of M Ver HL TOS Total Length Identification Flags Fragm. Offset Inner Header TTL Protocol Header Checksum Src Address = Addr of C An IP packet is received at the Home agent from a correspondent host for a mobile host. Dest Address = Addr of M IP PAYLOAD CS 515 © İbrahim Körpeoğlu

IP-IP Decapsulation at the Care-of-Address
Ver HL TOS Total Length Identification Flags Fragm. Offset Outer Header TTL Protocol=4 Header Checksum Src Address = Home agent addres An encapsulated IP packet is received at the foreign agent (or at the mobile Itself for a collocated care-of-address). Receiver understands that the packet is IP-IP encapsulated by looking to the protocol field (which is 4). Dest Address = Care-of-Address of M Ver HL TOS Total Length Identification Flags Fragm. Offset Inner Header TTL Protocol Header Checksum Src Address = Addr of C Dest Address = Addr of M IP PAYLOAD Receiver forwards (not routes) the decapsulated IP packet to the mobile node using link-level mechanisms! CS 515 © İbrahim Körpeoğlu

Minimal Encapsulation at Home Agent
Tunneled to care-of-address Ver HL TOS Total Length Identification Flags Fragm. Offset Outer header TTL Proto=55 Header Checksum Ver HL TOS Total Length Src Address = Addr of home agent Identification Flags Fragm. Offset Dest Address = Care-of-addr of mobile TTL Protocol Header Checksum Protocol S Reserved Header Checksum Src Address = Addr of C Minimal Inner header Src Address = Addr of C Dest Address = Addr of M Dest Address = Addr of M IP PAYLOAD IP PAYLOAD Encapsulated using Minimal Encapsulation Method Packet comes from Correspondent host CS 515 © İbrahim Körpeoğlu

Home Network Configurations
Physical Home Network Internetwork 1) Router Home Agent Internetwork Physical Home Network Router and home agent 2) Internetwork Virtual Home Network 3) Router and home agent CS 515 © İbrahim Körpeoğlu

Sending packets between mobile and foreign agent
When a mobile moves to a new location, a foreign should be broadcasting (IP and link layer broadcast) advertisements on the link (sub-network). Mobile will be able to receive this broadcast message and will learn: The IP address of the foreign agent (this will be the care-of-address of the mobile most of the time). The hardware (MAC or link-level address) of the foreign agent. When mobile sends a registration packet through this foreign agent, the foreign agent will learn: The home address of the mobile The hardware (MAC or link level) address of the mobile. The registration packet will be sent directly to the foreign agent by using the MAC address of the foreign agent (No need to do ARP request). CS 515 © İbrahim Körpeoğlu

FA periodically broadcasts advertisements.
Mobile Node - M Foreign Agent - FA Mobile Node receives broadcast frame and learns the MAC and IP address of the FA. Its Stored this info. Broadcasted Mobile Agent Advertisement FA periodically broadcasts advertisements. MAC broadcast address is used. No need for ARP. Mobile Node sends a registration request message directly to FA. It is not using ARP protocol to obtain the MAC address of FA. FA learns the MAC address of a mobile from the registration request message. Learns also the home address of the mobile. This info is stored. Registration Request Registration Reply Reply is sent directly to the MAC address of mobile. No need for ARP. DATA Mobile node sends data Directly to the MAC address of FA. No ARP needed. FA sends data directly to the MAC address of FA. No ARP needed. DATA CS 515 © İbrahim Körpeoğlu

Sending Data from Foreign Agent to Mobile
Node Foreign Agent APPS Other Fields UDP TCP/UDP Dst Src IP Payload IP_M IP_C …. IP_F IP_M MAC_F IP Payload IP_M IP_C …. type MAC_F MAC_M MAC_M Src (6 b ytes) Dst (6 bytes) IP Header Ethernet Header (link level header) CS 515 © İbrahim Körpeoğlu

Sending Data from Mobile to Foreign Agent
Node Foreign Agent APPS APPS Other Fields IP Payload TCP/UDP TCP/UDP Src Dst …. IP_M IP_C IP Payload IP_F IP_M MAC_F MAC_M type …. IP_M IP_C IP Payload MAC_F MAC_M Dst (6 bytes) Src (6 b ytes) IP Header Ethernet Header (link level header) CS 515 © İbrahim Körpeoğlu

Decapsulation again Mobile Node Foreign Agent APPS APPS Home Agent
TCP/UDP TCP/UDP dst src dst src ds t src IP_M IP_M IP_F IP_H IP_M IP_C IP_M IP_H IP_F TUNNEL MAC_F IP_M IP_C MAC_F MAC_M MAC_M CS 515 © İbrahim Körpeoğlu

How to attract packets at the Home network
Physical Home Network Proxy ARPing enabled MAC_R IP_M MAC_H Internetwork Router Proxy ARP table Home Agent IP Payload IP_M IP_C …. MAC_H An IP Packet comed from a correspondent host destined to a Mobile Host Broadcast ARP Request Who has IP_M Unicast ARP Reply I have IP_M, My MAC addr=MAC_H IP Packet put into a Ethernet Frame IP Payload IP_M IP_C type MAC_R MAC_M CS 515 © İbrahim Körpeoğlu

Proxy ARPing The packet comes to the last router that the home subnetwork is connected to. The router will try ro resolve the IP address of Mobile (IP_M) into the corresponding MAC layer address (Hardware address). For this pupose, it will broadcasts an ARP request packet Since the mobile is not at home subnet, it will not be able to answer ARP request. Home agent will answer instead of the Mobile node. İn order to do this, home agent should be configured to do proxy ARPing. Home agent replies to the ARP request with an ARP reply, including its MAC address (MAC_H) as the MAC level address corresponding to the IP address of the Mobile. The router, upon receiving the ARP reply, will send the IP packet to the MAC address of the home agent. In this way, the home agent attracts the IP packets that are destined to the mobile node. CS 515 © İbrahim Körpeoğlu

Gratuitous ARP Functionality
Physical Home Network Mobile Node is at home subnet ARP Table IP_M  MAC_M An Other Host MAC_R Internetwork ARP Table IP_M  MAC_M Router An Other Host Home Agent MAC_H Mobile Node MAC_M Mobile Node moved away from homesubnet ARP Table IP_M  MAC_H An Other Host MAC_R Internetwork ARP Table IP_M  MAC_H Router An Other Host Home Agent MAC_H Physical Home Network CS 515 © İbrahim Körpeoğlu

Gratuitous ARP Operation
An Other Host MAC_R Internetwork Router An Other Host Home Agent MAC_H Physical Home Network Home Agent Receives Registration Request from New Location Home agent broadcasts Gratuitous ARP on the Link (indicating IP_M is now located at MAC addr MAC_H) All other hosts on the LAN update their ARP Caches with binding: IP_M MAC_H CS 515 © İbrahim Körpeoğlu

ARP Packet Format Ether Type: 0x8006 ARP protocol
Op Field: – ARP Request – ARP Reply Ethernet Header ARP Packet Ether Dst Ether Src Ether Type Hw type Prot type op Sender Hw Addr Sender IP Addr Target Hw Addr Target IP Addr Hw size Proto size Sender Receiver LAN ARP Request (Broadcasted) ARP Reply (Unicasted) CS 515 © İbrahim Körpeoğlu

Example: Proxy ARP (IP_X, MAC_X) Src Dst Correspondent Host (IP_C)
(IP_H, MAC_H) Host X Home Agent ---- IP_C IP_M IP Payload Normal Internet Routing INTERNET Home Subnet Router (IP_R, MAC_R) ARP Request FFFFFF MAC_R Hw type Prot type 1 MAC_R IP_R ---- IP_M (IP_M, MAC_M) Sender MAC Sender IP Target MAC Target IP Proxy ARP Reply IP_M MAC_H IP_H MAC_H 2 Prot type Hw type MAC_H MAC_R Target IP Target MAC Data (IP Packet) MAC_H MAC_R ---- IP_C IP_M IP Payload CS 515 © İbrahim Körpeoğlu

Example: Gratuitous ARP
Correspondent Host (IP_C) (IP_X, MAC_X) (IP_H, MAC_H) Host X IP_M  MAC_M Home Agent IP_M  MAC_H INTERNET Home Subnet Router REGISTRATION (IP_R, MAC_R) IP_M  MAC_M (IP_M, MAC_M) IP_M  MAC_H Broadcast Gratuitous ARP Request (IP_M, MAC_M) IP_M ..... IP_M MAC_H 1 Prot type Hw type MAC_H FFFFFF Target IP Target MAC Sender MAC Sender IP Home Agent Broadcast an Gratuitous ARP Request on the LAN. Any receiveing host will update its ARP cache. CS 515 © İbrahim Körpeoğlu

Route Optimization in Mobile IP

Triangular Routing Internet 2 3 Foreign Agent Home Agent 4 Mobile Node
1 5 Forward and reverse paths are different. This causes triangular routing. Path fromc correspondent hosts to mobile hosts may not be optimal: All packets has to go through home agent. Correspondent Host CS 515 © İbrahim Körpeoğlu

Solution Approach Let the correspondent hosts know the current mobility binding or just binding (home address  care-of-address mapping) for mobile hosts. They will store this binding. They will use this binding to directly send the packets to the current location of the mobile. They will again use encapsulation since the care-of-address may not be always collocated at the mobile node (foreign agent should decapsulate). The encapsulated packets will go to the care-of-address directly without going through the home agent. Correspondent hosts should support the binding protocol: Need for modification at correspondent hosts!. CS 515 © İbrahim Körpeoğlu

Binding Update How does a correspondent host will learn the current binding for the mobile node? Let the mobile node inform the correspondent host! For example when it receives a packet from a correspondent host Let the home agent inform the correspondent host. This is the method chosen, since it is easier to establish security association between a home agent and a correspondent host (Binding update should be secure so the malicious users can not send binding updates to the corresspondent hosts without authenticating themselves). CS 515 © İbrahim Körpeoğlu

Binding Update INTERNET Dst Src IP Payload IP-M IP-C …. First Packet
Mobile Host C Dst Src IP Payload IP-M IP-C …. Home Agent H First Packet INTERNET Foreign Agent F Other Fields Dst Src IP Payload IP-M IP-C …. Other Packets Dst Src IP Header Dst Src Binding Update (IP-M, IP-C) IP Payload IP-M IP-C …. IP-F IP_C …. Inner IP Header Outer IP Header Correspondent Host C IP_MIP_F ..... Binding Cache Packet Transport from a Correspondent Host to a Mobile CS 515 © İbrahim Körpeoğlu

Binding Warning/Request
A correspondent host may request a binding Update message from Home agent. Correspondent host sends a Binding Request message and waits for a Binding Update Message. A mobile node may warn a Home agent (or some other agent) to send a Binding Update message to a particular host (a correspondent host or to some other host). Mobile node sends a Binding Warning message. Binding warning message include the host IP address (called target address field) to where an Update will be sent. A host receiving a Binding Update message should send back a Binding Acknowledgement message. The sender of Binding Update may retranmit Binding Update if it did not received a Binding Acknowledgement message. The retransmission should occur after a backoff time. All binding messages are sent over UDP. CS 515 © İbrahim Körpeoğlu

Smooth Handoffs CH HA FA3 Registration
INTERNET FA3 Registration Packets may be dropped during handoffs. Handoff to a new base station (or foreign agent) and registration with home agent takes time. During this time, packets will be forwarded to the old base station (FA), where the mobile node moved away from. FA1 FA2 M M CS 515 © İbrahim Körpeoğlu

Smooth Handoffs CH 1) When M makes handoff, it warns the new Foreign agent (FA2) (by use of a Binding Warning message) to send a Binding Update message to the old FA (FA1). HA INTERNET FA3 Registration FA1 Binding Update FA2 2) New FA sends a Binding Update message to old FA. 3) Old FA re-encapsulates the Ip packets received from home agent and sends them to the new FA. M M CS 515 © İbrahim Körpeoğlu

Supporting Fast Handoffs in Mobile IP

Fast Handoffs For highly mobile users, handoffs will be too frequent. Implications of this: Handoffs should be very fast in order to minimize packet delays and packet losses. Registration will be too frequent: Registration causes delay Registration causes extra signaling (control) traffic in the wireless link and infrastructure. Two solution approaches to support fast handoffs: Use of IP multicasting Use of hierarchical foreign agents. CS 515 © İbrahim Körpeoğlu

Use of IP Multicasting A collection of foreign agents in the vicinity of each other join to a multicast group. The group will have a multicast IP address. Mobile node will use this multicast IP address as the care-of-address. The home agent will send the encapsulated packets for the mobile to this multicast IP address. Foreign agents in the multicast group will buffer the received encapsulated IP packets for a while before discarding In this way, when a mobile handoffs from one FA to an other FA (in the same multicast group), it will be able to recover the packets transmitted during handoff from the new FA. CS 515 © İbrahim Körpeoğlu

Hierarchical Foreign Agents
Uses a hierarchy of foreign agents between mobile node and home agent. Aims is to localize handoffs and registration. The hierarchy could be consisting of for example: Base stations (access points) at the lowest level – leaf. Intermediate routers between base stations and campus edge routers in a campus. Campus edge router at the highest level (root) of the foreign agent hierarchy. CS 515 © İbrahim Körpeoğlu

Binding MHFA1 HA INTERNET MHFA2 FA1 MHFA3 MHFA5 MHFA6 MHFA4 FA2 FA3 FA4 FA5 FA6 FA7 MHIF MHIF MHIF MH MH MH CS 515 © İbrahim Körpeoğlu

The following functions of Mobile IP is enhanced: Agent Advertisements Registration Data Forwarding CS 515 © İbrahim Körpeoğlu

Agent Advertisements CH HA Mobility Agent Extension
to ICMP Router Advertisement INTERNET Type Length Sequence Number Lifetime Flags Reserved FA1 Zero or more care-of-addresses ………. FA1 FA1 FA2 FA3 Agent Advertisement message Care-of-Address field content FA2, FA1 FA2, FA1 FA3, FA1 In a message. FAx denotes the IP address of Foreign Agent X. FA4 FA5 FA6 FA7 FA4, FA2, FA1 FA6, FA3, FA1 MH CS 515 © İbrahim Körpeoğlu

Registration CH MHFA1 HA 0 8 16 31 Type Flags Lifetime MHFA2
Registration Request Format INTERNET Type Flags Lifetime MHFA2 Home address=MH FA1 Home agent=FA2 Care-of—address=FA5 MHFA4 FA2 MHFA5 FA3 Identification Extensions (Authentication Extension) ….. MHIF MHIF FA4 FA5 FA6 FA7 FA5, FA2, FA1 REG PKT MH MH FA4, FA2, FA1 FA4, FA2, FA1 Compare FA4, FA2, FA1 With FA5, FA2, FA1 CS 515 © İbrahim Körpeoğlu

Forwarding Src Dst CH CH MH IP payload IP Packet MHFA1 HA
Src Dst Src Dst Dst INTERNET FA1 HA CH MH IP payload Encapsulated IP Packet FA1 MHFA2 Src Dst Src Dst FA1 FA2 CH MH IP payload MHFA5 FA2 FA3 Each FA takes an encapsulated packet from previous FA (or HA) and recapsulates the packet to be sent to the next FA. If an FA is the final FA on the way to the mobile node, then it does not recapsulate the packet. FA2 FA5 CH MH IP payload FA4 FA5 MHIF FA6 FA7 CH MH IP payload MH CS 515 © İbrahim Körpeoğlu

Cellular IP İbrahim Körpeoğlu

Motivation Mobile IP can work for any link type:
Ethernet, Token Ring, Wireless LAN (802.11), Bluetooth, PPP, etc. This implies different types of mobility Slow moving: between Ethernet links Fast moving: between wireless LAN access points Indoor: inside a building Campus-wide Wide-area: between campuses/sites. Mobile IP envisions handoff rates less than one registration per second. There is a need to support higher rate handoffs. Need for fast-handoffs, low packet delay, minimum packet losses Need for minimum mobility signaling (registration packets). CS 515 © İbrahim Körpeoğlu

Problems with Mobile IP
Registration takes time: Distance between home and foreign agents could be too large. Incurs packet delays and jitter Registration incurs extra load on Resource scarce wireless access links (air interface) Between mobile node and foreign agent On internet infrastructure (core network) Between foreign agent and home agent. Mobile IP causes registration overhead even the mobile is not sending or receiving data while it is moving. Need for labeling a mobile node as in active or passive mode. CS 515 © İbrahim Körpeoğlu

Cellular IP Approach Use the concept of cellular mobile telephone networks for Handoff management Efficient handoffs with low delay, minimum packet losses. Location tracking Exact location is known for active mobiles Approximate location is known for passive mobiles Paging is used to learn the exact location for a passive mobile. Passive connectivity Based on IP principles: The underlying network is IP. No new packet formats No encapsulation No new address space. CS 515 © İbrahim Körpeoğlu

Cellular IP Cellular IP can support micro-mobility in
Pico-cellular or micro-cellular networks (Personal Area Networks or Wireless LANs) Campus wide networks Multi-cell wireless access networks. Can be integrated with Mobile IP to support macro-mobility Mobility between campuses and different administrative domains. CS 515 © İbrahim Körpeoğlu

Cellular IP In a cellular IP network Related Work:
All routing for mobile hosts is done by Cellular IP routing Route distribution and update is done according to Cellular IP protocol. No need to modify the IP packet format or IP forwarding mechanism. Per-host location information is stored in cellular IP network routers for mobile hosts. Related Work: Hierarchical Foreign Agents for Mobile IP proposed by IETF Hawaii Project at Lucent/Bell Labs Learning features of Ethernet switches A switch learns the location of traffic sources while its is forwarding the frames. CS 515 © İbrahim Körpeoğlu

Cellular IP Network Model
CH HA Mobile IP enabled INTERNET GW/FA Foreign Agent (Care-of-address) Gateway Router Cellular IP Access Network BS BS Base station BS BS MH Mobile Host CS 515 © İbrahim Körpeoğlu

Beacons CH HA GW: IP address of Gateway
Mobile IP enabled INTERNET GW: IP address of Gateway BSx: IP address of base-station X MH: Home IP address of Mobile GW Gateway Router Beacons are sent periodically originating from Gateway Beacons GW GW BS1 BS2 GWGW GWGW Pointer to Gateway BS2 BS2 BS3 BS4 GWBS2 GWBS2 BS4 Beacons are use to let the routers learn the path to the gateway. MH Mobile Host CS 515 © İbrahim Körpeoğlu

Data Transport From MH To CH CH HA MH CH IP pyld
Mobile IP enabled INTERNET MH CH IP pyld GW: IP address of Gateway BSx: IP address of base-station X MH: Home IP address of Mobile Route Cache Entry for Mobile GW Gateway Router MHBS2 Route Cache Entry For GW MH CH IP pyld BS1 BS2 GWGW GWGW MHBS3 MH CH IP pyld BS3 BS4 GWBS2 MHIF GWBS2 MH CH IP pyld Src dst MH Mobile Host (Active) CS 515 © İbrahim Körpeoğlu

Data Transport From CH To MH Src Dst CH CH MH IP pyld HA
Mobile IP enabled INTERNET Src Dst Src Dst GW: IP address of Gateway BSx: IP address of base-station X MH: Home IP address of Mobile HA GW CH MH IP pyld Route Cache Entry for Mobile GW Gateway Router MHBS2 Route Cache Entry For GW CH MH IP pyld BS1 BS2 GWGW GWGW MHBS3 CH MH IP pyld BS3 BS4 GWBS2 MHIF GWBS2 CH MH IP pyld src dst MH Mobile Host (Active) CS 515 © İbrahim Körpeoğlu

Route Updates CH HA GW: IP address of Gateway
Mobile IP enabled INTERNET GW: IP address of Gateway BSx: IP address of base-station X MH: Home IP address of Mobile Route Cache Entry for Mobile GW Gateway Router MHBS2 Route Entry Refreshed! Route Cache Entry For GW MH GW ICMP BS1 BS2 GWGW GWGW MHBS3 Route Entry Installed! Route Entry Timeout!!! MH GW ICMP BS3 BS4 GWBS2 MHIF GWBS2 Route Entry Installed! Route Update messages are ICMP messages with special type and code fields. MH GW ICMP Src dst Route Update message generated by mobile!!! MH Mobile Host (Active) CS 515 © İbrahim Körpeoğlu

Handoff CH HA GW: IP address of Gateway
BSx: IP address of base-station X MH: Home IP address of Mobile Mobile IP enabled INTERNET Route Cache Entry for Mobile GW Gateway Router MHBS2 Route Cache Entry For GW RT Update MHBS3 MHBS4 BS1 BS2 GWGW GWGW Route Update Packet RT Update GWBS2 GWBS2 RT Update BS3 BS4 MHIF MHIF RT Update MH MH CS 515 © İbrahim Körpeoğlu

Paging CH HA GW: IP address of Gateway
BSx: IP address of base-station X MH: Home IP address of Mobile Mobile IP enabled INTERNET Gateway Router Route Cache Entry for Mobile GW MHBS2 MHBS2 Paging Cache BS7 Page MH Page MH Page Update messages are ICMP messages with special type and code fields. Page MH BS1 BS2 BS6 MHBS3 Page MH Page MH BS8 Page MH BS5 BS3 BS4 MHIF MHIF Page Update Packet Page MH Route Update Packet Data Packet Mobile Host (Passive) MH Page Packet CS 515 © İbrahim Körpeoğlu

BlueSky: A new indoor wireless access network and mobility solution
İbrahim Körpeoğlu CS 515 – Mobile and Wireless Networking Bilkent University Computer Engineering Department Bilkent/Ankara CS 515 © İbrahim Körpeoğlu

Putting the pieces together
BlueSky is an experimental protoype wireless access network designed at IBM T.J. Watson Research Center, Hawthorne, New York, in 1998. It has been started before Bluetooth Special Interested group (SIG) has been formed. In this course, we are now at a level to understand the design of wireless systems from bottom to nearly top. CS 515 © İbrahim Körpeoğlu

Why BlueSky? Aims providing wireless indoor access for palmtop computer (PDAa) such as palm pilots. Low power and low cost are primary objectives. No service fee. is expensive, power hungry Cellular modems have expensive service charge. Aims providing indoor seamless mobility support for Bluesky enabled devices and computers. Least modification to end-devices with different operating systems. PPP constrained. CS 515 © İbrahim Körpeoğlu

Why networking for PDAs?
There are different kinds of palmtop computers. Also called personal digital assistants. Examples Palm Pilots: running PalmOS Pocket PC: running Windows CE We concentrated on a solution for Palm Pilots. CS 515 © İbrahim Körpeoğlu

Palm Pilot Networking at that time
Desktop PC Palm Pilot Ethernet Card Serial Port Serial Port (UART) RS-232C Serial Cable Internet Apps Apps TCP TCP IP IP PPP PPP Ethernet Driver Serial Port (UART) Serial Port (UART 16550) Ethernet Card Internet RS-232C Serial Cable Ethernet Cable CS 515 © İbrahim Körpeoğlu

BlueSky Wireless Access Design
Palm Pilot Serial Communication over Serial Port Access Point Ethernet Card Serial Port Serial Port Internet Wireless Adapter Wireless Modem Apps Telnet Access Point Palm Pilot TCP TCP IP IP Optional PPP PPP Ethernet Driver Radio Device(FSK) Radio Device(FSK) Ethernet Card Ethernet Cable Internet Radio device is a wireless modem that takes serial stream of bytes as input and modulates them over the 915Mhz carrier frequency using FSK modulation. CS 515 © İbrahim Körpeoğlu

BlueSky Wireless Channel - Downlink
There are two wireless physical channels that are used between an access point and mobile stations. Downlink channel (channel 1 – f1) Uplink channel (channel 2 – f2) Namely. FDD scheme is used for duplexing. Palm Pilot 1 Serial Port Wireless Adapter Access Point Ethernet Card Palm Pilot 2 Serial Port Wireless Modem Downlink Channel (Channel 1) Serial Port Wireless Adapter When two palm-pilots transmit at the same time, collisions may occur on the uplink channel. Hence, we need Media Access Control for Uplink. CS 515 © İbrahim Körpeoğlu

BlueSky Wireless Channel - Uplink
Uplink channel is shared between mobiles that are in the control of the access point. Since it is shared collisions may occur without media access control. Palm Pilot 1 Uplink Channel (Channel 2) Serial Port Wireless Adapter Access Point Ethernet Card Palm Pilot 2 Collision! Serial Port Wireless Modem Serial Port Wireless Adapter When two palm-pilots transmit at the same time, collisions may occur on the uplink channel. Hence, we need Media Access Control for Uplink. CS 515 © İbrahim Körpeoğlu

BlueSky Wireless Channel Access
Palm Pilot 1 The uplink channel is access by using a polling based media access control protocol. Serial Port Wireless Adapter Access Point Ethernet Card Palm Pilot 2 Serial Port Wireless Modem Serial Port Wireless Adapter When two palm-pilots transmit at the same time, collisions may occur on the uplink channel. Hence, we need Media Access Control for Uplink. CS 515 © İbrahim Körpeoğlu

BlueSky Wireless Access Design with MAC
Palm Pilot Serial Communication over Serial Port Pseudo-Serial Communication over Virtual Port Access Point Ethernet Card Serial Port Serial Port Internet Wireless Adapter Wireless Modem Apps Telnet Access Point Palm Pilot TCP TCP IP IP Optional PPP PPP Ethernet Driver Wireless Adapter Wireless MAC Wireless MAC Ethernet Card Ethernet Cable Internet Radio Device(FSK) Radio Device(FSK) CS 515 © İbrahim Körpeoğlu

Wireless Adapter Design
Palm Pilot Apps TCP IP IP Hdtr IP Payload PPP PPP Hdr IP Hdr IP Payload CRC UART Driver stream of bytes Serial Port of PDA (UART) Serial Port (UART1655) Wireless Adapter i386 CPU MAC Protocol MAC Hdr PPP Hdr IP Hdr IP Payload CRC FCS MS-DOS stream of bytes PC Board Serial Port Radio Transceiver (FSK Modulation) CS 515 © İbrahim Körpeoğlu

Wireless Adapter – Outside view
Wireless adapter envisioned product form factor. Wireless adapter working prototype (little bit big!!) CS 515 © İbrahim Körpeoğlu

connected to radio transceiver
Wireless Adapter – Inside – Development Board Parallel Port connected to radio transceiver PC-104 Board A/C Adapter Serial port Connected to radio transceiver Serial port connected to PDA Antenna i386 CPU 2-4MB RAM Flash Disk to store files (2MB) Radio Transceiver RSSI Pin CS 515 © İbrahim Körpeoğlu

Wireless Adapter – MAC Development Cycle
PC with windows Write MAC protocol software for mobile wireless adapter (using C in Visual C++ environment) in a standalone PC with windows. Compile the software for DOS and obtain mac.exe 1 Transfer the MAC executable to development board using a floppy (dev. board has a floppy drive). At the dev. board, copy the executable from floppy to flash disk of each wireless adapter. 2 Monitor FLASH DISK Repeat 1,2,3,4 until Success! 3 Access Point Physically transfer the flash disks to wireless adapters (be careful not to break the pins!) Dev. Board 4 Run the prototypes and test! CS 515 © İbrahim Körpeoğlu

10Base-T Ethernet Card (optional) Floppy Disk Drive (optional)
What is PC104 Basically a PC with different form factor (shape and size) Stackable cards Video cards, ethernet cards, flopply drive, ... Cards are connected originally with 104 pins. Cards are much smaller than ISA cards used in PCs. Suitable for embedded systems. PC-104 Board VGA Card (optional) 10Base-T Ethernet Card (optional) Floppy Disk Drive (optional) CS 515 © İbrahim Körpeoğlu

RS-232C Standard for transmitting serial stream of data over short-distances. - Between two computers in a room - Between a computer and a modem Between a computer and a device Between two devices Defines: - The signal levels to represent bits/bytes The line types in a cable to transmit/receive data and control signals. DTE Device DCE Device RS232C cable Telephone line modem DTE Device DTE Device DTE: Data Terminal Equipment DCE: Data Commmunication Equipment (Circuit terminating) RS232C cable CS 515 © İbrahim Körpeoğlu

UART: Serial Port Interface IP COMPUTER PPP Operating System
UART Driver Microprocessor Interface UART: Universal Asynchronous Receiver Transmitter Serial Port Interface (UART) UART Core Rx FIFO (16 bytes) Tx FIFO (16 bytes) UART 16550 RS232 Cable CS 515 © İbrahim Körpeoğlu

RS232C Connectors – 25 Pin DB25 Male Connector DB25 Female Connector
TxD: Transmite Data (from DTE to DCE) RxD: Receive Data (from DCE to DTE) RTS: Request to Send (from DTE to DCE) CTS: Clear to Send (from DCE to DTE) DSR: Data Set Ready (from DCE to DTE) DTR: Data Terminal Ready (from DTE to DCE) SD: Signal Ground CD: Carriier Detect (from DCE to DTE) 1 13 14 25 TxD 2 12 15 24 Tx D 3 11 16 23 RTS 4 10 17 22 RI CTS 5 9 18 21 DSR 6 CD 8 19 20 DTR SG 7 SD 7 20 DTR 19 1 DSR CD 8 6 1 21 18 CTS 9 5 1 22 17 RTS 10 4 23 16 Rx D 11 3 24 15 12 Tx D 2 25 14 13 1 CS 515 © İbrahim Körpeoğlu

RS232C Connectors – 9 Pin DB9 Male Connector DB9 Female Connector SG 5
1 9 6 DSR DTR 4 RxD 2 8 CTS 7 RTS TxD 3 TxD 3 7 RTS 8 CTS RxD 2 DTR 4 6 DSR 9 1 SG 5 TxD: Transmite Data (from DTE to DCE) RxD: Receive Data (from DCE to DTE) RTS: Request to Send (from DTE to DCE) CTS: Clear to Send (from DCE to DTE) DSR: Data Set Ready (from DCE to DTE) DTR: Data Terminal Ready (from DTE to DCE) SD: Signal Ground CD: Carriier Detect (from DCE to DTE) CS 515 © İbrahim Körpeoğlu

RS232C Data Transmission +15V LSB MSB 0 1 0 0 0 0 0 1 0 1 1 Space (=0)
Space (=0) +3V Intermediate Region 0V -3V Mark (=1) 7 data bits Start bit Parity bit -15V Two stop bits Data byte transmission corresponding to ascii character A (65) CS 515 © İbrahim Körpeoğlu

Connnecting Terminals using RS-232
DTE Device Computer DCE Device MODEM 2 TxD TxD 2 3 RxD RxD 3 4 RTS RTX 4 5 CTS CTS 5 Female connector Male connector DTE Device Computer DTE Device Computer 2 TxD TxD 2 3 RxD RxD 3 4 RTS RTX 4 5 CTS CTS 5 Female connector Null-Modem Cable Female connector CS 515 © İbrahim Körpeoğlu

MAC Protocol Polling based Works over full-duplex wireless channel
Uplink channel: from mobiles to access point (shared) Downlink channel: from access point to mobiles (no concention) Support sleep modes to save power In its simplest form, round-robin poll schedule With appropriate scheduding can support different rates to each mobile MAC Hdr (8 bytes) Data (optional) (0-2K bytes) CRC (or FCS) (2 bytes) MAC Header Fields: Temp MAC address: 1 bytes Length: 2 bytes MAC start frame delibiator Type of Frame: POLL, NODATA, DATA, REGISTRATION, etc Flags Sequence No: 1 bit ACK (1 bit) CS 515 © İbrahim Körpeoğlu

Radio Channels over which MAC operates (2 channels)
full duplex MAC MAC radioAP radio1 Access Point MAC radio2 Ad an ON/OFF switch to each radio and control it using a MAC protocl MAC radio3 CS 515 © İbrahim Körpeoğlu

MAC protocol overview MAC MAC MAC MAC MAC Poll-R1 radioAP radio1 DATA
Access Point Poll-R2 No-DATA Invite MAC Poll-R3 Register radio2 DATA MAC MAC radio4 radio3 CS 515 © İbrahim Körpeoğlu

MAC Protocol Concept - 1 AP-MAC: AP side of the MAC protocol
M-MAC: mobile side of MAC protocol M1 M-MAC RADIO AP-MAC M2 M-MAC M1 .... Data FCS AP RADIO RADIO Channel 1 M3 M-MAC RADIO MAC Header Payload CRC Addr .... Data FCS DATA is variable size with some maximum value. CS 515 © İbrahim Körpeoğlu

MAC Protocol Concept - 2 M1 M-MAC FCS Data .... AP RADIO Channel 2
AP-MAC M2 M-MAC AP RADIO RADIO M3 M-MAC RADIO CS 515 © İbrahim Körpeoğlu

MAC Protocol Concept - 3 M1 M-MAC RADIO AP-MAC M2 M-MAC M2 .... FCS AP
POLL RADIO RADIO Channel 1 M3 M-MAC RADIO CS 515 © İbrahim Körpeoğlu

MAC Protocol Concept - 4 M1 M-MAC RADIO AP-MAC M2 M-MAC FCS Data ....
Channel 2 M3 M-MAC RADIO CS 515 © İbrahim Körpeoğlu

MAC Protocol Concept - 5 M1 M-MAC RADIO M-MAC M3 .... Data FCS AP-MAC
Channel 1 M3 M-MAC RADIO DATA is variable size with some maximum value. CS 515 © İbrahim Körpeoğlu

MAC Protocol Concept - 6 M1 M-MAC RADIO AP-MAC M2 M-MAC AP RADIO RADIO
FCS .... AP RADIO NODATA Channel 2 CS 515 © İbrahim Körpeoğlu

MAC Registration and change of point of Attachment.
Both AP1 and AP2 are using the same uplink and downlink radio channels. Therefore, they should not overlap. This problem can be solved by using more channels. AP1 AP2 INVITE 2 RPL REG 3 1 M1 M2 M3 M3 M4 M5 M3 moves out of region of AP1. After some threshold time it will detect that is lost connection will start listening to INVITE packets from other access points if there exists any. When it enter to the region of AP2, it will hear an INTIVE from AP2 and will responde with a REGISTRATION REQUEST after some random backoff time. AP2 will reply to REGISTRATİON REQUEST with a REPLY. REPLY will contain a temporary 1 byte MAC address for M3. M3 will receive REPLY and now it is registrated with AP2. Data Communication can resume over AP2. AP1 will delete M3 from its table after some period. CS 515 © İbrahim Körpeoğlu

Pseudo-Serial Communication over Virtual Port
BlueSky Access Point Design Pseudo-Serial Communication over Virtual Port Access Point Problem Multiplex traffic (PPP packets) from multiple PPP interfaces at the MAC layer and send them as MAC payloads to appropriate mobile devices depending on which PPP interfaces the packets has came from. Demultiplex MAC frame payloads to different PPP interfaces depending on from which mobiles the frames came from (again at the MAC layer) Ethernet Card Serial Port Internet Wireless Modem Telnet Access Point TCP Run multiple PPP Interfaces for each mobile device registered. IP Optional PPP PPP PPP Ethernet Driver Wireless MAC Ethernet Card Ethernet Cable Internet Radio Device(FSK) CS 515 © İbrahim Körpeoğlu

BlueSky Access Point Design
Internet Ethernet Cable Radio Transceiver Serial Port Serial Cable Serial Port Access Point is Unix Box connected to a radio transceiver over a serial cable. It can run BSDI/OS or Linux. It may have a monitor connected for debugging purposes. MAC protocol is implemented as a user level process that talks to the serial port (hence to radio transceiver) on one end and to the PPP interfaces at the other end. Access point can be configured as IP forwarder to forward the IP packets back and forth between Internet (or LAN) and Wireless Link Access point terminates the PPP connections from mobile device (namely AP runs a PPP server for each mobile that is running a PPP client and that is connected to the AP). CS 515 © İbrahim Körpeoğlu

BlueSky Access Point Implementation
Telnet Server Process AP MAC Process MAC Protocol /dev/ttya /dev/pty1 /dev/pty2 /dev/ptyn TCP CRC IP Pkt PPP_H Unix Kernel IP Pkt Routing Table IP FCS CRC IP Pkt PPP_H MAC_H Network Interfaces PPP0 PPP1 PPPn Ether0 Ethernet Driver UART Driver pty1 pty2 ptyn tty1 tty2 ttyn CRC IP Pkt Eth Hdr UART Ethernet Card Pseudo-terminal Masters Pseudo-terminal Slaves Ethernet Cable Radio Device(FSK) Frame Coming from Mobile 2 Frame going to LAN CS 515 © İbrahim Körpeoğlu

What is PPP? PPP is Point-to-point protocol
It is a protocol to encapsulate IP datagrams so that they can be sent over serial lines Serial line transmission is just a stream of bytes. It does not have any packet boundaries. Connection oriented. A connection has to be established between PPP peers (PPP client and PPP server) before transmitting data (IP packets most of the time). PPP functions Framing Encapsulation of Network Layer packets into PPP frames so that they can be sent over serial lines Link control protocol To establish, test, and configure data-link connections. Network control protocol To configure the network layer protocols. CS 515 © İbrahim Körpeoğlu

PPP Framing 1 1 1 2 up to 1500 bytes 2 1 Flag 0x7E Addr 0xFF Control
Protocol Information CRC Flag 0x7E PPP Frame format (used for both control and data packets) 0x0021 IP packet PPP DATA PACKET LCP: Link Control Protocol NCP: Network Control ACP: Authentication Control Protocol PPP CONTROL PACKET 0xc021 LCP Data 0x8021 NCP Data PPP CONTROL PACKET CRC is computed over Addr, Control, Protocol, and Information fields before byte stuffing. CS 515 © İbrahim Körpeoğlu

Why PPP Framing needed? APP TCP ……………… IP IP Length IP Payload Length
Assume we run IP directly over Serial Port without using PPP. TCP ……………… IP IP Length IP Payload Length IP Payload Length IP Payload Serial Port Driver Serial Port Driver Ethernet Driver Serial Port (UART) Serial Port (UART 16550) Ethernet Card RS-232C Serial Cable IP Packets are sent over serial line without any other encapsulation. Serial line is not error-free. Some bytes can get corrupted. Assume the receiver IP detects the boundaries of IP packets by looking to the length field in the header of IP packets. If the byte(s) that is corrupted carries the length field of the IP packet that is currently transmitted over the serial line, the the receiving IP layer can not detect the end-of-IP-packet correctly. Hence we will loose synchronization between two ends and no further IP packet can be received correctly at the receiver side. CS 515 © İbrahim Körpeoğlu

Byte Stuffing Replace the occurrence of 0x7e with 0x7d, 0x5e
Replace the occurrence of 0x7d with 0x7d, 0x5d Hdr Data CRC ff 03 00 21 56 99 01 7e 05 7d 17 7d 7e 89 19 7d 7e ff 03 00 21 56 99 01 7d 5e 05 7d 5d 17 7d 5d 7d 5e 89 19 7d 5d 7e Byte Stuffed Content Start of Frame End of Frame 0x7e is the flag byte. 0x7d is the escape byte. Inside the frame, no occurrences of 0x7e after byte stuffing. 0x7e is in hexadecimal. It is equal to in binary. CS 515 © İbrahim Körpeoğlu

PPP Functions Apps TCP IP IP PPP Connection Establishment PPP
Client Side 1) PPP Server Side Ethernet Driver Serial Port (UART) Serial Port (UART 16550) Ethernet Card 2) Data Transfer Ethernet Cable PPP Connection Terminationt 3) Link Establishment Authentication Network Configuration Data Transfer Termination PPP Connection Establishment PPP Connection Termination CS 515 © İbrahim Körpeoğlu

PPP Connection Establishment
Established Established P C L I E N T P S E R V Set IP Address Set DNS Address Provide IP Address, DNS Address, etc. NCP (CONF_REQ) 3 NCP (CONF_ACK) Provide Info for Authentication (uname, password) ACP (REQ) Authenticate Using PAP, CHAP EAP etc. 2 ACP (ACK) Provide incoming MTU Set auth protocol to be used. Set outgoing MTU LCP Packet (CON_REQ) Set outgoing MTU Provide incoming MTU Provide which Auth protocol to be used. 1 LCP Packet (CONF_ACKL) CS 515 © İbrahim Körpeoğlu

PPP Connection Establishment
After PPP connection establishment A PPP client learns at least: the IP and DNS addresses that it will use to connect to the Internet the link MTU (maximum PPP packet size) that is to be exported to the IP layer, so that IP layer fragments IP packets that are larger than the MTU PPP Connection establishment Takes time! Incurs a lot of control traffic (control packets) over the low bit-rate serial (wired or wireless) link. CS 515 © İbrahim Körpeoğlu

Our wireless link and PPP
Access Point IP /24 eth0 Apps RT Table pp1 TCP pp2 Mobile A IP ( ) PPP1 PPP2 Ethernet Driver PPP Connection PPP AP Wireless MAC Ethernet Card Wireless MAC Radio Device(FSK) Radio Device(FSK) Apps PPP Connection TCP IP Mobile B PPP Wireless MAC Radio Device(FSK) CS 515 © İbrahim Körpeoğlu

Looking back to our solution
Our solution so far provides Wireless connectivity for multiple mobile devices that are connected to a single access point. By use of PPP, mobile devices can run IP and related applications and access point can act as a forwarder. We may have more than one access point installed in a target region. We will mobile connected to some access point. However, the solution so far does not provide seamless mobility between access points, since PPP connection will break upon change of access point and higher layer protocols and applications will be disturbed. CS 515 © İbrahim Körpeoğlu

MAC level registration Wireless MAC AP Wireless MAC Ethernet Card
Access Point A Apps IP RT Table TCP /24 eth0 Data Flow pp1 Mobile A IP ( ) PPP Connection PPP PPP1 Ethernet Driver MAC level registration Wireless MAC AP Wireless MAC Ethernet Card carrier detection Radio Radio moves Mobile’s IP address changes! Router Access Point B Apps IP RT Table TCP /24 eth0 pp1 Mobile A IP ( ) PPP Connection PPP PPP1 Ethernet Driver MAC level registration Wireless MAC AP Wireless MAC Ethernet Card carrier detection Radio Radio CS 515 © İbrahim Körpeoğlu

BlueSky’s Approach Do not terminate PPP at the access point.
Means: do not run PPP server at the access point. Run PPP at a backend server. Tunnel PPP traffic from access points to the backend server. Establish PPP once Redirect PPP traffic to a new tunnel when point of access changes. CS 515 © İbrahim Körpeoğlu

PPP Backend Server (PBS)
Concept PPP Backend Server (PBS) IP PPP Eth INTERNET PBS LAN Data Flow Tunnel 2 (PPP/UDP/IP/LAN) Tunnel 1 LAN LAN AP1 PPP Connection AP2 PPP PPP M1 M2 M3 M3 M4 M5 Mobile’s IP Address remains same. CS 515 © İbrahim Körpeoğlu

Protocol Architecture
Palm Pilot Apps PPP Backend Server TCP IP IP PPP Access Point PPP Tunnel Tunnel Agent UDP UDP Tunnel Agent Wireless MAC Wireless MAC IP IP Radio Device(FSK) Radio Device(FSK) Ethernet Ethernet Router IP Pyld IP H IP Pyld IP H CRC IP Pyld IP H PPP_H CRC IP Pyld IP Hdr PPP_H FCS CRC IP Pyld IP H PPP_H MAC_H CRC IP Pyld IP H PPP_H CRC IP Pyld IP Hdr PPP_H TUN_H CRC IP Pyld IP H PPP_H TUN_H CRC IP Pyld IP H PPP_H TUN_H UDP_H IP_H CS 515 © İbrahim Körpeoğlu

AP_MAC/Tunnel Process
BlueSky New Access Point Implementation AP_MAC/Tunnel Process CRC IP Pkt PPP_H MAC Protocol Tunnel Protocol FCS CRC IP Pkt PPP_H MAC_H CRC IP Pkt PPP_H TUN_H /dev/ttya UDP socket Unix Kernel UDP IP UART Driver Ethernet Driver UART Ethernet Card Wireless channel Radio Device(FSK) LAN CS 515 © İbrahim Körpeoğlu

BlueSky PPP Backend Server Implementation
Tunnel Protocol PPP_H IP Pkt CRC TUN_H PPP_H IP Pkt CRC /dev/pty1 /dev/pty2 /dev/ptyn Tunnel Process UDP socket UDP UDP_H TUN_H PPP_H IP Pkt CRC IP Pkt Routing Table Network Interfaces PPP0 PPP1 PPPn Ethernet Interface Ethernet Driver pty1 pty2 ptyn tty1 tty2 ttyn CRC IP Pkt PPP_H Ethernet Card Pseudo-terminal Masters Pseudo-terminal Slaves From Access Point To Internet Frame going to LAN Common Data path CS 515 © İbrahim Körpeoğlu

Advantages of terminating PPP at the backend server and BlueSky in general
Authentication is easier. Password files can be kept at a central location (PBS) Otherwise they should be replicated in each access point. PPP connection is established once for data communication It is not terminated while changing point of attachment. Only terminated when mobile no longer wants to access the network IP address remains same although point of access changes. IP layers is unaware of mobility (hence all other higher layers: TCP, UDP, Aplications) Can support fast handoffs (PPP connection establishment takes time). Is more bandwidth efficient for the resource scarce wireless access link. Total seamless mobility support Mobility can be supported in this way for other network layer protocols other than IP such as IPX and Appletalk. No need to modify mobile devices hardware and software. No need to modify different OS stacks for PDAs. CS 515 © İbrahim Körpeoğlu

Limitations of terminating PPP at the backend server and BlueSky in general.
Solutions is restricted with PPP as the data link layer. All traffic has to go through PPP backend server This can be optimized by parsing PPP packets at the access point and looking to the destination IP address The bandwidth of the tunnel path (10Mbps) is much much higher than the bandwidth of the wireless link (150 Kbps) Packet delays are negligible over the wired links. Running PPP between IP and packet oriented MAC layer introduces: Extra framing (one at PPP layer, one at the MAC layer) Extra connection time: MAC registration + PPP connection establishment. CS 515 © İbrahim Körpeoğlu

BlueSky Summary Wireless Access Mobility Shared 150 Kbps bandwidth
FSK modulation at 915MHz 10m range Full duplex channel (FDD duplexing) Polling based access-point-coordinated MAC protocol Mobility Solution at the data-link layer Works for PPP and various network protocols: IP, IPX, AppleTalk Roaming in a single administrative domain. CS 515 © İbrahim Körpeoğlu

BlueSky versus Bluetooth
Developed by a small group of people at IBM in 1998. Developed by a consortium consisting of 5 big companies (IBM, Intel, Toshiba, Ericsson, Nokia) and more than 500 adaptor companies. Started in May 1998. 10m range FSK modulation GFSK modulation 915 MHz operating frequency 2.4 GHz operating frequency FDD duplexing TDD duplexing Access point coordinated polling based MAC (no strict time slotting) Master coordinated polling based MAC based on strict time slotting. 150 Kbps shared bandwidth (no limit on devices other than performance). 1Mbps shared bandwidth (at most 8 devices: 1 master, 7 slaves) Complete networking solution. Does not define exactly how IP can be run over it. Complete mobility solution. Does not define a mobility solution. CS 515 © İbrahim Körpeoğlu

Mobile and Wireless Networking Ibrahim Korpeoglu

Similar presentations

Presentation on theme: "Mobile and Wireless Networking Ibrahim Korpeoglu"— Presentation transcript:

Similar presentations

About project

Feedback

Log in

Auth with social network:

Mobile and Wireless Networking Ibrahim Korpeoglu

Similar presentations

Presentation on theme: "Mobile and Wireless Networking Ibrahim Korpeoglu"— Presentation transcript:

Similar presentations

About project

Feedback