A Comparison of Wi-Fi and Wimax with Case Studies Ming-Chieh Wu

Florida State University Libraries

Electronic Theses, Treatises and Dissertations The Graduate School

2007 A Comparison of Wi-Fi and WiMAX with Case Studies Ming-Chieh Wu

Follow this and additional works at the FSU Digital Library. For more information, please contact [email protected]

THE FLORIDA STATE UNIVERSITY

COLLEGE OF ENGINEERING

A COMPARISON OF WI-FI AND WIMAX WITH CASE STUDIES

Ming-Chieh Wu

A Thesis submitted to the Department of Electrical and Computer Engineering in partial fulfillment of the requirements for the degree of Master of Science

Degree Awarded: Fall Semester, 2007 The members of the Committee approve the Thesis of Ming-Chieh Wu defended on October 30th, 2007.

Bruce A. Harvey Professor Directing Thesis

Ming Yu Committee Member

Simon Y. Foo Committee Member

Approved:

Victor DeBrunner, Chair, Department of Electrical and Computer Engineering

Ching-Jen Chen, Dean, FAMU-FSU College of Engineering

The Office of Graduate Studies has verified and approved the above named committee members.

TABLE OF CONTENTS List of Tables…………………………..………….………………………………………..……vi List of Figures………………………………….…………………………………………….…vii List of Abbreviations……………………..………….………………………………………….ix ABSTRACT……………………………………………..……………………………………...xii CHAPTER ONE 1. Introduction …………………………………………………………………………………....1 CHAPTER TWO 2. 3G…………………………………..…………………………………………………………..5 2.1. Introduction………………………………………………………………………...…..5 2.2. Technologies…………………………………………………………………………...5 2.2.1. WCDMA……………………………….………………………………………..5 2.2.2. CDMA2000………………………………………………………………….…..6 2.2.3. TD-SCDMA……………………………………………………………….…….6 2.3. Development…………………………………………………..……………………….7 2.4. Cases………………………………………………………………..…………….…….8 2.5. Conclusion……………………………………………………………………….……..9 CHAPTER THREE 3. IEEE 802.11, Wireless LAN…………………………………………………...……………..10 3.1. The background of IEEE 802.11…………………………………………….………10 3.2. Capacity……………………………………………………………………….….….11 3.3. The Physical Layer and MAC Layer…………………………………………...……14 3.3.1. The Physical Layer………………………………………………..……….….14 3.3.1.1. Introduction……………………………………………………….……14 3.3.1.2. PLCP and PMD…………………..………………………...……….……15 3.3.1.3. CS/CCA………………………………………………………….……….15 3.3.1.4. IEEE 802.11b, DSSS and HR-DSSS…………………………..…...…..16 3.3.1.4.1. Theory and Transmission method…………………………………16 3.3.1.4.2. Against interference……………………..…………………………17 3.3.1.4.3. PLCP and PMD of DSSS……………………...... ……..17 3.3.1.4.4. HR-DSSS and CCK…………………………………………..……18 3.3.1.4.5. PLCP and PMD of HR-DSSS………..……………....……………19 3.3.1.5. IEEE 802.11a and OFDM…………………………………………….…20 3.3.1.5.1. Background…………………………………………………..…….20 3.3.1.5.2. Principles…………………………………………………………..20 3.3.1.5.3. The IEEE 802.11a’s OFDM and its PLCP and PMD……………..22 3.3.1.6. IEEE 802.11g and the Physical Layer……………………………………24 3.3.1.6.1. Background and Backward Compatibility………….……...……24 3.3.1.6.2. ERP-OFDM and DSSS-OFDM………………………………..…..25 3.3.2. MAC Layer………………………………………………………………..…….27 3.3.2.1. Introduction……………………………………………...….……………27 3.3.2.2. DCF and PCF…………………………….……………………………..27

iii

3.3.2.3. Hidden node and CSMA/CA……………………………...... ……………28 3.3.2.4. Fragmentation………………………………………………………….30 3.3.2.5. Interframe spacing………………………………………………………..30 3.3.2.6. Power saving………………………………..…………………………….31 3.3.2.7. Security…………………………………….……………………………..32 3.4. The next generation standard…………………………………………………………..34 3.4.1. IEEE 802.11n…………………………………………..………………………..34 3.4.2. MIMO (Multiple-Input/Multiple-Output)………………….………………...…35 3.5. Limitation………………………………………………………..……………………..35 CHAPTER FOUR 4. IEEE 802.16, Wireless MAN………………………………………………………...……….37 4.1. The background of IEEE 802.16……………………………………...….………….37 4.2. Capacity of the IEEE 802.16 family………………………………………...…………39 4.2.1. IEEE 802.16………………………………………………………………….….39 4.2.2. IEEE 802.16a………………………………………………………………..…..39 4.2.3. IEEE 802.16c…………………………………………………………..………..39 4.2.4. IEEE 802.16-2004.……………………………………………...…...………….39 4.2.5. IEEE 802.16e-2005…………………………………………………………..….40 4.2.6. IEEE 802.16f, IEEE 802.16g and IEEE 802.16h……………….…………….40 4.3. The Physical Layer and MAC Layer……………………………...... …...……………42 4.3.1. IEEE 802.16-2004……………………………………………….………..…..42 4.3.1.1. Physical Layer…………………………………………….…………....42 4.3.1.1.1. Four transmission mode: SC, SCa, OFDM and OFDMA…………44 4.3.1.1.2. AMC (Adaptive Modulation and Coding)….…………………...44 4.3.1.1.3. Channel-quality Measurement……………………………………..45 4.3.1.1.4. OFDM in WiMax……………………...………….……………….45 4.3.1.2. MAC Layer……………………………………………………….…..…..46 4.3.1.2.1. Three sublayers: CS, CPS and SS…………...…….………………46 4.3.1.2.2. AAS (Advanced Antenna Systems)………………….…...……..47 4.3.1.2.3. QoS……………………………………………………..………….47 4.3.1.2.4. Security…………………………………………………………….49 4.3.1.2.4.1. Overview……………………………………...……………49 4.3.1.2.4.2. Overall Analysis………………….……………...………..52 4.3.2. IEEE 802.16e-2005, the newest standard…………………………………….....53 4.3.2.1. OFDMA-PHY………………………………...………………………..53 4.3.2.1.1. Background…………………..…………………………………….53 4.3.2.1.2. Frame structure…………………….………………………………53 4.3.2.1.3. Various Subcarrier Allocation Modes……………………………..54 4.3.2.2. MAC Layer…………..…………………………………………………..56 4.3.2.2.1. QoS…………………….…………………………………………..56 4.3.2.2.2. Power-saving………….………………………………………...…57 4.3.2.2.2.1. Sleep Mode…………….……………………………………57 4.3.2.2.2.2. Idle Mode……...…………………………………………….58 4.4. The future of WiMax…………………………….……………..……………………58 CHAPTER FIVE

5. Case study…………………………………………………………………………………….61 5.1. FSU wireless network………………………………………………………………….61 5.1.1. Introduction………………………………………………………………...……61 5.1.2. Implementation…………….……………………………………………………62 5.1.3. Analysis…………………………………………………………….…………73 5.1.4. Conclusion…………….……………………………………...…………………74 5.2. WiFly in Taipei……………………………..………………………………………….74 5.2.1. Introduction………………………..…………………………………………….74 5.2.2. Implementation………………………………………………………………….75 5.2.3. Analysis…………………………....………………………………………….83 5.2.4. Conclusion……………………….…………………………………………....86 5.3. Wibro in South Korea…………………………….……………………………………86 5.3.1. Introduction…………………………….………………………………………..86 5.3.2. History…………………………………………………………………….…..87 5.3.3. Implementation………………………………………………...... …..88 5.3.4. Analysis……………………………………………………………………...….92 5.3.5. Conclusion……………………………..…………………….……………….....93 5.4 . Overall analysis………………………...………………………..……………..………94 5.5. Other developing wireless network………...……………………………….……….95 CHAPTER SIX 6. Conclusion…………………………………………………..…………………….………..98 BIBLIOGRAPHY……………………………………...…………………………….………..101 BIOGRAPHICAL SKETCH…………………………………………………………………110

LIST OF TABLES Table 1-1 802.3 Family……………………………………………………..…………..……….2 Table 3-1 Family of IEEE 802.11……………..……..……………………..…………………11 Table 3-2 Capacities for 802.11a/b/g……………….……...…………….………………………12 Table 3-3 IEEE 802.11b/g Channel Used for Different Countries……….……...………………12 Table 3-4 IEEE 802.11a Channel Use for North America…………….……………...... 13 Table 3-5 OFDM Modulation and Data Rate…………………...…………………..………..23 Table 4-1 The Basic Data of IEEE 802.16………………………………………...….41 Table 4-2 Modulation and Coding Supported in WiMax……………..…………..…..45 Table 4-3 OFDM Parameters Used for WiMAX…………………..…………..…….46 Table 4-4 Service Flows in WiMax……………………...……………………..…….48 Table 4-5 DL Distributed Subcarrier Permutation (FUSC)………….………..………55 Table 4-6 UL-DL Adjacent Subcarrier Permutation ………………………….………….……..56 Table 5-1 The Basic Information of FSUWIN………………………………………...…66 Table 5-2 Xirrus XS-3700 AP Technology Data………….……………………………....66 Table 5-3 Foundry Networks IP250 Technology Data……………...……….……...... ….66 Table 5-4 Vivato 2.4 GHz Indoor & Outdoor Wi-Fi Switch Technology Data…...…..…68 Table 5-5 Parameters of WLAN and Mesh Network……………………….....………..70 Table 5-6 Features of Nortel Wireless Mesh Network………………..….…………..77 Table 5-7 The Numbers of Subscribers………………...……………..…………………79 Table 5-8 The Comparison of WiBro and WiMax……………………...….……………..92 Table 5-9 The Basic Data of FSUWIN, WiFly and WiBro…………..…………………96 Table 5-10 Current Internet Access Technologies……………….……..…………..……..97 Table 6-1 3G, Wi-Fi and WiMax Overall Comparison………………………….………..98

LIST OF FIGURES Figure 2-1 The Evolution of 3G………………………………...………...…………..7 Figure 3-1 OSI Model………………………………………..…………..…………..14 Figure 3-2 802.11 PLCP, PMD and MAC Structure…………..………………..……15 Figure 3-3 DSSS Transmission………………………………..…………………...…16 Figure 3-4 DSSS PLCP………………………………………..………………..……18 Figure 3-5 CCK Modulation…………………………………..……………………..19 Figure 3-6 HR/DSSS PLCP Framing…………………………..…………………….19 Figure 3-7 FDM and OFDM…………………………………..……………………..20 Figure 3-8 CP and PP………………………………………..…….…………………21 Figure 3-9 OFDM Modem……………………………………..….…………………22 Figure 3-10 Constellations Diagram…………………………..……………………..23 Figure 3-11 OFDM PLCP Structure………………………….………………………24 Figure 3-12 DSSS-OFDM PSDU Format………………………..…..………………26 Figure 3-13 DSSS-OFDM Long Preamble Structure……………..…………………..26 Figure 3-14 DSSS-OFDM Short Preamble Structure……………………..…………..27 Figure 3-15 The Hidden Node Problem………………………..………...……………28 Figure 3-16 CDMA/CA………………………………………..…………….………29 Figure 3-17 Virtual Channel Sensing…………………………..……………………..29 Figure 3-18 A Fragment Burst…………………………………..…………………....30 Figure 3-19 Frame Interval for IEEE 802.11…………...………..…………………..31 Figure 3-20 WEP Frame and Operation…………………………..…………………..33 Figure 3-21 The WEP Encryption Process………………………….………………..33 Figure 4-1 The Layer Structure of IEEE802.16……………………..………...………43 Figure 4-2 A Spatial Multiplexing MIMO System…………………....………………44 Figure 4-3 PKM Authorization Process and Parameters………………..…………….50 Figure 4-4 PKM Protocol Messages Exchange Process and Parameters…….....……………..51 Figure 4-5 Encryption Frame Structure and Process……………………..…….……..52 Figure 4-6 OFDMA Frame Structure……………………………………..…………..54 Figure 4-7 Coverage and Capacity for Different Wireless Access Techniques…..….....59 Figure 5-1 The Coverage of FSU Wireless Network………….…..……………………….62 Figure 5-2 Fisher Lecture Hall………………………..………..…………………….64 Figure 5-3 Landis Green………………………………………..…..………………..65 Figure 5-4 Shores Library……………………………………..………………….….65 Figure 5-5 Stadium……………………………………………..……………………66 Figure 5-6 Xirrus XS-3700 AP…………………………………………………………..67 Figure 5-7 Foundry Networks IP250……..………………………..………...…………..69 Figure 5-8 Vivato 2.4 GHz Indoor & Outdoor Wi-Fi Switches………..……………..72 Figure 5-9 FSUWIN Login System……………………………………..…………….74 Figure 5-10 Wireless Mesh Network Structure I……………...………..…………….76 Figure 5-11 Wireless Mesh Network Structure II………………….…..…………….76 Figure 5-12 Wireless Access Point 7220………………………………..….………..77 Figure 5-13 Nortel Wireless Mesh Network Solution Example…………..…………..80 Figure 5-14 The Relationship Between WiMax and WiBro……...……..……………..87 Figure 5-15 The Network Structure of WiBro……………………..……..……………87

vii

Figure 5-16 The Operation Band of WiBro……………………………………..………...…….88 Figure 5-17 WiBro Features………………………...…………………..…………..……89 Figure 5-18 MAC Layer Model…………………………………………..……….……..90 Figure 5-19 U-RAS Premium……………………………………………..…….………91 Figure 5-20 ACR Basic Data……………………..……………………………………….91 Figure 5-21 An Example Application of 3G, Wi-Fi and WiMax…………………..……..95 Figure 6-1 Cooperation of WiMax and Wi-Fi………………………………………...…100

viii

LIST OF ABBREVIATIONS

Abbreviation Full Name Page 3G Third generation of mobile phone standards and technologies 5 3GPP 3rd Generation Partnership Project 5 AAS Advanced Antenna Systems 42 ACK Acknowledgement 26 AGC Automatic Gain Control 21 AMC Adaptive Modulation and Coding 40 AP Access Point 32 APEC Asia Pacific Economic Cooperation 93 ARPA Advanced Research Projects Agency 1 BE Best-effort service 44 BER Bit Error Rate 20 BS Base Station 29 BTC Block Turbo Codes 40 BWA Broadband Wireless Access 34 CCK Complementary Code Keying 16 CDMA/CA Carrier Sense Multiple Access/Collision Avoidance 25 CDMA/CD Carrier Sense Multiple Access/Collision Detection 26 CDMA2000 1X Evolution-Data Only 6 EV-DO CID Connection Identifier 42 COFDM Coded OFDM 20 CP cyclic prefix 19 CS/CCA Carrier Sense/Clear Channel Assessment 14 CT Communication Technology ix CTC Convolution Turbo Codes 40 CTS Clear To Send 23, 26 DBPSK Differential Binary Phase shift Keying 15 DCF Distributed Coordination Function 25 DECT Digital-Enhanced Cordless Telephony 34 DIFS DCF InterFrame Spacing 28 DL Downlink 48 DQPSK Differential Quadrature Phase shift Keying 15 DSSS Direct Sequence Spread Spectrum 14 EIFS Extended InterFrame Spacing 28 ERP Extended Rate PHY 22 ertPS Extended Real-Time Polling Service 51 FCC Federal Communications Commission 11 FDD Frequency Division Duplexing 35 FDM Frequency Division Multiplexing 18 FDMA Frequency Division Multiple Access 47 FEC Forward Error Correction 40 FFT Fast Fourier Transform 19

FOMA Freedom of Mobile Multimedia Access 8 FSUWIN Florida State University Wireless Integrated Network 55 FUSC Fully Used Subchannelization 48, 49 GPRS General Packet Radio Service 5 Global System for Mobile Communications/Pan-European digital GSM 5 cellular land mobile telecommunication system ICI inter-carrier interference 19 ICV Integrity Check Value 29 IEEE Institute of Electrical and Electronic Engineers 1 IFFT inverse Fast Fourier Transform 19 IMT International Mobile Telecommunication 5 ISI Inter-symbol interference 19 ISP Internet Service Provider 33 IT Information Technology ix ITU International Telecommunication Union 5 IV Initialization Vector 29 KT Korean Telecoms Industry 86 LDPC Low Density Parity Check 40 LMDS Local Multipoint Distribution Systems 34 LOS Line of Sight 34 MAC Media Access Control 25 MIMO Multiple-Input/Multiple-Output 31 MMDS Multichannel Multipoint Distribution Services 34 MMS Multimedia Message Service) 5 MPDUs MAC Protocol Data Units 13 MS Mobile Station 28 MSDUs MAC Service Data Units 41 NAV Network Allocation Vector 23, 26 NLOS Non-Line of Sight 35 nrtPS Non-real-time polling service 43 OFDM Orthogonal Frequency Division Multiplexing 9, 19 OFDMA Orthogonal Frequency Division Multiple Access 47 OSI Open System Interconnection 12 OTC Office of Telecommunications and Networking 55 PBCC Packet Binary Convolution Coding 22 PCF Point Coordination Function 25 PIFS PCF InterFrame Spacing 28 PKM Privacy and Key Management 44 PLCP Physical Layer Convergence Protocol 13 PMD Physical Medium Dependent 13 PN Codes Pseudorandom Noise Codes 15 PP cyclic postfix 19 PPDU PLCP Protocol Data Unit 13 PRNG Pseudorandom Number Generator 29 PSDU PLCP Service Data Unit 13

PSS Portable Subscriber Station 89 PUSC Partially Used Subchannelization 48, 50 QAM Quadrature Amplitude Modulation 20 QoS Quality of Service 42 RAS Radio Access Station 89 RCF Request to Send 26 RS-CC Reed Solomon – Convolution Code 40 rtPS Real-time polling services 43 SA Security Associations 44 SDUs Service Data Units 51 SIFS Short Inter-Frame Space 23, 28 SKT South Korea Telecom 8 SS Subscriber Station 44 STBC Space-Time Block Coding 32 TCP/IP Transmission Control protocol/Internet Protocol 3 TDD Time Division Duplexing 35 TDM Time Division Multiplexing 35 TD-SCDMA Time Division - Synchronized Code Division Multiple Access 5 UGS Unsolicited grant services 43 UL Uplink 48 UMTS Universal Mobile Telecommunications System 5 VoIP Voice over IP 32 WCDMA Wideband Code Division Multiple Access 5 WEP Wired Equivalent Privacy 29 WiBro Wireless Broadband Access Service 86 Wi-Fi Wireless fidelity 31 WiMax World Interoperability for Microwave Access 35 WISP Wireless Internet Service Provider 34 WLAN Wireless Local Area Network 9 WLL Wireless Local-Loop 34 WMAN Wireless Metropolitan Area Network 35 WWW World Wide Web 3

ABSTRACT Currently over 50% of the world’s population check their e-mails everyday. Collecting information from the Internet is a routine. In the early 21st century, wireless communication has become a hot topic in IT (Information Technology) and CT (Communication Technology), as evidenced by the growth of wireless technologies such as 3G, Wi-Fi and WiMax. 3G is a cellular technology developed in conjunction with the cellular phone network. Wi-Fi is a wireless local area network technology. WiMax is designed for the wireless metropolitan area network. Today, people not only want the fixed wireless access to the Internet, but also want the mobile wireless access as well. They want a ubiquitous connection, even when in a train, a cab, or the subway. This demand is resulting in increasing competition between the leading wireless technologies. 3G, Wi-Fi and WiMax all appear to have the potential to feed the demand, but still have issues that need to be addressed. The future direction of wireless Internet access is uncertain, including whether these three technologies will operate cooperatively or competitively. This thesis is going to predict the future direction by analysis of 3G, Wi-Fi and WiMax technologies and the evaluation of three wireless access case studies. This thesis will begin with an introduction to the history of Internet and will then continue with a discussion of the technical aspects of 3G, Wi-Fi and WiMax. After the technology introduction, this thesis will evaluate three current implementations of wireless Internet access as case studies to verify the capabilities of Wi-Fi and WiMax, and to discuss the feasibility of building a city-wide wireless network. Finally, a reasonable prediction of the future implementation of a city-wide wireless Internet structure will be presented.

xii

CHAPTER ONE

1. Introduction

The cellphone company Nokia uses the slogan of “connecting people” in its advertising campaign. That means that technologies should be based on demand and humility. In the field of the Internet, the basis of all research and innovation is geared towards making transmission of communication more efficient. The network technology that would eventually evolve into “the Internet” was initially developed by the military requisition called ARPANET, during late 1960s. ARPA is the acronym for Advanced Research Projects Agency an agency of the Department of Defense. This network was the pioneer of the “packet switch” type network. Little did the originators of ARPANET realize that this network would eventually lead to a method that could connect every PC in the world for sharing and exchanging information. The two most significant features of ARPANET were the concept of network routing and the use of packet for data transfer. In ARPANET, each computer was a node and it received the data and then routed it to next node. ARPANET was also the first network using packet switching to transmit data. Each node could send each packet to its destination by different paths, so it could improve the transmission to become more efficient and more reliable. In the beginning, this project only included four locations, Stanford Research institute, UCLA, UC Santa Barbara and Utah University. After it was initiated successfully in 1969, the locations extended to the east coast the following year including MIT, Harvard, Beranek and Newman. This was the world’s first WAN (Wide Area Network).

In 1973, the Xerox Corporation, located in California, developed the LAN (Local Area Network) for connecting PCs within a local area. Therefore, the Ethernet was formed. The DIX alliance (DEC, Intel and Xerox) was the first pusher of Ethernet and then transferred the patent right to IEEE (Institute of Electrical and Electronic Engineers). This move made Ethernet become popular very quickly. In 1982, DIX published Ethernet Version 2 (EV2) and later on IEEE published IEEE 802.3 CDMA/CD standard which was based on EV2 in 1983. Today, the IEEE 802.3 series is the most well-known Ethernet standard. The IEEE 802.3 is a big family that is presented in table 1-1[67][86]. They can support speeds from 10 Mbps to 1 Gbps. The increased technology causes the Internet to continually speed up.

Communication technology continues to develop and an increasing number of communication technologies have been implemented. However these technologies may not be compatible with each other. The ARPANET researchers developed a set of protocols called TCP/IP (Transmission Control protocol/Internet Protocol) to integrate networks that may be based on differing communications technologies and protocols. Internet research was not only conducted in U.S., but also in other places such as Europe, Asia, Canada and so on. Competing

1 protocols such as the X.25 protocol from Europe were developed, but eventually TCI/IP became accepted as a world-wide standard and the world-wide Internet was formed.

Table 1-1 802.3 Family

Ethernet Date Description Standard Experimental 1972 2.94 Mbit/s (367 kB/s) over coaxial cable (coax) cable bus Ethernet 10 Mbit/s (1.25 MB/s) over thin coax (thinnet) - Frames have a Type field. This Ethernet II 1982 frame format is used on all forms of Ethernet by protocols in the Internet (DIX v2.0) protocol suite. 10BASE5 10 Mbit/s (1.25MB/s) over thick coax - same as DIX except Type IEEE 802.3 1983 field is replaced by Length, and an 802.2 LLC header follows the 802.3 header 802.3a 1985 10BASE2 10 Mbit/s (1.25 MB/s) over thin Coax (thinnet or cheapernet) 802.3b 1985 10BROAD36 802.3c 1985 10 Mbit/s (1.25 MB/s) repeater specs 802.3d 1987 FOIRL (Fiber-Optic Inter-Repeater Link) 802.3e 1987 1BASE5 or StarLAN 802.3i 1990 10BASE-T 10 Mbit/s (1.25 MB/s) over twisted pair 802.3j 1993 10BASE-F 10 Mbit/s (1.25 MB/s) over Fiber-Optic 100BASE-TX, 100BASE-T4, 100BASE-FX Fast Ethernet at 100 Mbit/s (12.5 802.3u 1995 MB/s) w/auto negotiation Full Duplex and flow control; also incorporates DIX framing, so there's no 802.3x 1997 longer a DIX/802.3 split 802.3y 1998 100BASE-T2 100 Mbit/s (12.5 MB/s) over low quality twisted pair 802.3z 1998 1000BASE-X Gbit/s Ethernet over Fiber-Optic at 1 Gbit/s (125 MB/s) 802.3-1998 1998 A revision of base standard incorporating the above amendments and errata 802.3ab 1999 1000BASE-T Gbit/s Ethernet over twisted pair at 1 Gbit/s (125 MB/s) Max frame size extended to 1522 bytes (to allow "Q-tag") The Q-tag includes 802.3ac 1998 802.1Q VLAN information and 802.1p priority information. 802.3ad 2000 Link aggregation for parallel links 802.3-2002 2002 A revision of base standard incorporating the three prior amendments and errata 10 Gbit/s (1,250 MB/s) Ethernet over fiber; 10GBASE-SR, 10GBASE-LR, 802.3ae 2003 10GBASE-ER, 10GBASE-SW, 10GBASE-LW, 10GBASE-EW 802.3af 2003 Power over Ethernet 802.3ah 2004 Ethernet in the First Mile 802.3ak 2004 10GBASE-CX4 10 Gbit/s (1,250 MB/s) Ethernet over twin-axial cable 802.3-2005 2005 A revision of base standard incorporating the four prior amendments and errata.

Table 1-1 Cont.

802.3an 2006 10GBASE-T 10 Gbit/s (1,250 MB/s) Ethernet over unshielded twisted pair(UTP) Backplane Ethernet (1 and 10 Gbit/s (125 and 1,250 MB/s) over printed circuit 802.3ap 2007 boards) 802.3aq 2006 10GBASE-LRM 10 Gbit/s (1,250 MB/s) Ethernet over multimode fiber On 802.3ar Congestion management Hold 802.3as 2006 Frame expansion exp. 802.3at Power over Ethernet enhancements 2008 802.3au 2006 Isolation requirements for Power Over Ethernet (802.3-2005/Cor 1) exp. 802.3av 10 Gbit/s EPON 2009 Fixed an equation in the publication of 10GBASE-T (released as 802.3-2005/Cor 802.3aw 2007 2) exp 802.3ax Move Link aggregation out of 802.3 to IEEE 802.1 2008 exp 802.3ay Maintenance to base standard 2008 Higher Speed Study Group. 40 Gb/s over 1m backplane, 10m Cu cable assembly exp. 802.3ba (several pairs) and 100m of MMF and 100 Gb/s up to 10m or Cu cable assembly, 2009 100 m of MMF or 40 km of SMF respectively

For creating a common information sharing space to physicists, Tim Berners-Lee (1955- , British) designed the World Wide Web (WWW). At that time he was in the European Laboratory for Particle Physics, CERN (Conseil Européen pour la Recherche Nucléaire, European Council for Nuclear Research), and wanted to create a simple way to share and cooperate with other physicists around the world. Now WWW becomes another name for the Internet.

In the 1990s, data transmission went into the wireless era. Wireless transmission includes Bluetooth, Infrared, RF, IEEE 802.11, IEEE 802.16 and 3G. Currently there are several competing communication technologies for providing wireless Internet access. The primary competitors are 3G, Wi-Fi (IEEE 802.11) and WiMax (IEEE 802.16). 3G is a cellular technology and is currently evolving into all-IP network. Wi-Fi is a wireless local area network technology designed for home and small area implementations. WiMax is a wireless metropolitan area network technology which can cover larger area and support mobile Internet access at speeds up to 120km/hr. Each individual technology is designed for a particular application, but their capabilities at least partially overlap. WiMax is a newer technology that promises longer ranges and mobile access. Wi-Fi has been used for ten years, but has recently been implemented in campus-wide and city-wide networks. Moreover 3G is trying to increase its market share in

Internet access using the wide availability of the cellular telephone networks. These factors make it really difficult to predict the future of wireless Internet access.

In this thesis, the focus is primarily implementations of IEEE 802.11 and IEEE 802.16. In chapter 2, 3G will be discussed briefly, but the low data rate of current 3G implementations limit it’s applicability for general wireless Internet access. And for clarifying the capabilities of Wi-Fi and WiMax, this thesis will include three case studies. The case studies are the FSU (Florida State Univ.) campus Wi-Fi network, the city-wide Wi-Fi network in Taipei, Taiwan, and the city-wide WiMax network in Seoul, Korea. By comparing these three cases, the design problems and operation difficulties can be identified. Taipei and Seoul are both the first complete city- wide wireless network in the world. The current statuses of their networks reflect the capabilities of WiMax and Wi-Fi. By study these case studies, this paper is going to evaluate the feasibility of building a Wireless Networked City using WiMax and/or Wi-Fi.

CHAPTER TWO

2. 3G

2.1. Introduction

3G stands for “third generation of mobile phone standards and technologies”. It was developed under the IMT-2000 program (IMT, International Mobile Telecommunication) by the International Telecommunication Union (ITU). 3G has three standards, WCDMA (Wideband Code Division Multiple Access), CDMA2000 and TD-SCDMA (Time Division - Synchronized Code Division Multiple Access). Also, 3G’s network is a wide area cellular telephone network. It can provide internet access and video telephone. Compared to 2G, the 3G has higher bandwidth and faster speed. With this advantage, 3G can provide varieties of service. 3G supports both fixed and mobile environment and is also backward compatible with 2G. [68][69][136] – [138]

When doing vocal transmission, 3G can use Circuit Switch Mode for voice and video phone; for internet access/data transmission, 3G uses Packet Switch Mode. In this mode, users pay for how much they used, for example 0.099 cents per packet. Multimedia Message Service (MMS) is an important application of 3G. Although MMS has been applied to General Packet Radio Service (GPRS) as the 2.5G major application, with 3G’s advantages, wireless communication companies can provide more choices (video, audio, pictures and text) with MMS.

2.2. Technologies 2.2.1. WCDMA

WCDMA was developed by an organization called 3GPP (3rd Generation Partnership Project, December 1998) and also a part of IMT-2000 program. The WCDMA is based on GSM (Global System for Mobile Communications/Pan-European digital cellular land mobile telecommunication system) and GPRS. GSM is based on Circuit Switch Mode and GPRS is based on Packet Switch Mode. In Europe, WCDMA is called Universal Mobile Telecommunications System (UMTS). This technology came from the early third generation wireless network studies in Japan and Europe. Since the GSM was such a success, Europe began actively working on developing 3G. In 1995, Europe established Advanced Communications Technologies and Services (ACTS) for research and development of 3G. Also, in Japan the Association of Radio Industries and Businesses (ARIB) was established in 1993 as a committee for studying and developing Japan’s 3G technology. The WCDMA has several versions, R99, R4, R5 and the following ones. [69][139] – [141]

R99 was introduced in 1999 and could support 64 kbps circuit switching payload and up to 2Mbps packet switching payload. It is also compatible with GSM and GPRS services. The core network structure has two layers, Circuit Switch Layer and Packet Switch Layer. Circuit Switch Layer used GSM network and Packet Switch Layer used GPRS network. R99 inherited services and features of GSM and GPRS. The commercialized WCDMA network in the world are all based on this WCDMA Release 99 version. [69]

R4 was finalized in March 2001. Comparing with R99, the network structure did not change. The most different parts were in interface setup and functions improvements. One major change was in MSC (Mobile Switching Center). R4 split it into two parts MSC service and MGW (Media Gateway). The idea behind this was to separate control and payload. The other major change was introducing IP transmission to Circuit Switch Layer. R4 is an important step to All IP network. [69]

R5 was stopped upgrading in September 2002. It is a starting point of all-IP network of WCDMA network. R5 supports IP transmission and services. The most important component of R5 was IMS (IP Multimedia Subsystem). This subsystem can provide multimedia service based on the Internet. It is the bridge to integrate cellular system and the Internet. [69]

2.2.2. CDMA2000

CDMA2000 is the trade mark of TIA-USA (Telecommunication Industry Association) and which developed by 3GPP2. There were several versions of CDMA, CDMA2000 1X, CDMA2000 1X EV-DO. [69][142][143]

The evolution direction was the same with WCDMA, All IP network. Therefore, in CDMA- LMSD (Legacy MS Domain), the separation of control and payload was introduced. There are four phases in the CDMA2000 evolution. In phase 0, the main issues were to improve the present system. In phase 1/2 were to upgrade the system to become the All IP network system. In phase 3, the All IP network was formed and MMD (Multimedia Domain) started to show in the system. [69]

2.2.3. TD-SCDMA

This standard was individually developed by China and co-work with 3GPP. Future development is still ongoing. [69] [144]

2.3. Development

3G has been in development for couple years already. During this period of time, the telecommunication industry almost crashed in Europe. 3G is a promising technology, but the over-expectation made the European telecommunication industries spend too much money on the license of band operation, set up infrastructures and marketing. The result was that many companies cannot recoup their profits, and many of them had to give up the license and projects in order to lower the debt or litigate with government. These facts caused the development of 3G slowed down greatly in Europe. The most successful case in Asia is Japan. There are two companies running the business, NTT DoCoMo and KDDI. The NTT DoCoMo uses WCDMA and KDDI uses CDMA2000. Taiwan initialed the 3G service in 2005. In North America 3G is in sprout stage. Around the world, there are total 29 countries has 3G network so far. Figure 2-1 shows the evolution steps of 3G. [69][77]

Figure 2-1 The Evolution of 3G

The popularization of 3G has several factors, such as the user’s habits, market positioning, applications, and infrastructures. 3G has two opponents which are Wi-Fi and WiMax. They could be partners or enemies and it all depends on applications and market positioning. These wireless standards have their own advantages. 3G has bigger coverage and support mobility, but has lower data rate, and Wi-Fi has higher data rate, but smaller coverage and does not support

7 mobility. On the other hand, WiMax has the biggest coverage, highest data rate and also support mobility. [69][78]

The applications of 3G partially overlap with Wi-Fi and WiMax. 3G is a mobile phone technology, so its main stage is in the cell-phone market; internet access is only an additional benefit. 3G’s signal will fade out fast when there are too many people in the nearby area tring to access the Internet at the same time. 3G’s interface is mainly onthrough the cell-phone. Considering the operation convenience, it is tolerable for checking or sending e-mail, time tables, or MMS. However for high throughput applications, 3G appears to be not as good as the other two standards. The cell-phone’s inborn limitation, keyboard and screen size, makes it unsuitable for internet access. Moreover, 3G’s data rate and bandwidth are not fast enough for internet use.

2.4. Cases

Japan is a successful case for 3G development and mobile Internet access and there are some unique reasons for this. Japanese 3G service was started in late 2001 called FOMA (Freedom of Mobile Multimedia Access) by NTT DoCoMo. FOMA is the brand name of NTT DoCoMo’s 3G service. As the network coverage rate became higher and higher, most Japanese started to get used to checking and sending e-mail as well as getting daily information by phones. This phenomenon especially happens during rush hours in mass transit systems. Teenagers also like to share secrets, pictures and gossip by e-mail and MMS. Therefore, FOMA has become part of life and Multimedia service is the major applications of 3G. The most popular service in Japan now is called i-mode, it was in 2G and then upgrade to 3G. [69][76] – [78]

The 3G service in Korea is called JUNE. It is pushed by SKT (South Korea Telecom) in November 2002 and based on CDMA2000 1X EV-DO. Similar to Japan, it is also focused on multimedia service. In August 2003, just 8 month later, the users already increased to one million. JUNE has a music channel, video channel, MMS service and NOUL (Korean Idol). It is the first service which realizes the commercial video phone-call. [69]

In Taiwan, 3G cell-phones have become the main stream and the system also has upgraded completely. However, because of Wi-Fi and WiMax, telecommunication industries are still seeking where the business and balance point are. It is a true struggle in the communication market. Varieties of websites and software fit users’ different demands on the Internet. But with cell-phones, internet service is only provided by telecommunication industries and is very limited and not universal. In addition, because the cell-phones’ specifications are not unified, such as in the revolution of screen resolution and numbers of buttons, it is more difficult to design web pages as well as to promote the service. In Japan, the cell-phones are more unified, so it is easier to popularize mobile Internet access service. [69][76] – [78]

2.5. Conclusion

WAP (Wireless Application Protocol) is the previous model of mobile Internet access service, but it did not become as popular as “i-mode”. Poor contents and costly usage rate are two major reasons. 3G may revive it again, but it is an important lesson about the relationship between people and mobile Internet access service. Do people really need to use cell-phones to access the Internet? This is one of the key points for 3G’s future. And then after the final goal of 3G network is done, the All IP network, what will be the impact of the tele-market will be an important observation point.

Video phone-call is the major selling point of 3G. However there is a problem: who wants to be seen on thea phone call? Who wants to be awakened when the boss calls in the morning and be seen on the phone? Privacy is very important issue for everyone, but video phone-call has potential risk to break it. Therefore, it may take a long time for people to get used to it. In Taiwan, it has taken five years to get people used to the idea of pulling over to make a phone call and talking on the street. Currently, 99.9% population in Taiwan have a cell-phone. 3G still has a long way to go, because it takes time to change life styles and in most places in the world 3G has just arrived.

CHAPTER THREE

3. IEEE 802.11, Wireless LAN

3.1. The background of IEEE 802.11

When, IEEE published the standard 802.11Wireless Local Area Network protocol in 1997, it caused a revolution in methods of communication. People started to get rid of the constraint of wires and began to enjoy the freedom of unlimited space for getting information. It changed the way that people entertain themselves, get information, communicate and so on. IEEE 802.11 has many members in its family, and there are some of them that are widely used such as 802.11 a, 802.11b, and 802.11g. The newest incoming standard is 802.11n. The other name for 802.11 is called Wi-Fi. [01] - [03][22]

The first practical standard in the 802.11 family is 802.11b and it came out in 1999. The earlier standard 802.11 was too slow. Due to its sluggish speed, users could not experience the advantages of wireless communication. However, 802.11b was still very successful. Its speed may be only 5.5 Mbps and 11 Mbps [04]-[06], but it was fast enough to satisfy the basic requirement for browsing the Internet. In the past two years, this standard has become extremely popular. As a result, a PC which can support this protocol becomes a fundamental requirement.

Approximately around the same time when 802.11b came out, IEEE published another standard, 802.11a. These two standards used different operation bands and were not compatible. [07][08] 802.11a can reach the speed 54 Mbps and has more channels in its band to improve the bandwidth capacity for users. However, because of higher cost and the shorter transmission distance, this standard was only used by certain groups. [09][15]

The standard 802.11g was developed based on 802.11b and combined with OFDM (Orthogonal Frequency Division Multiplexing, used in 802.11a). Theoretically, its speed could reach 54 Mbps. This standard was published in the summer of 2003.. [10][20] Compared to the previous two standards, it was faster than 802.11b and had longer transmission distance and lower cost than 802.11a. It had the advantages of 802.11a and had backward compatibility with 802.11b. [09][11][15]

In the business field, rebuilding is always the last choice. 802.11b has existed in the market for a long time making it unwise to replace 802.11b with 802.11g. Therefore, the best method is to produce a device which can support both standards. Because they are compatible and operate in the same band, this is a more cost efficient option. For example Intel’s “Centrino” and “Sonoma” uses this type of production. [12] – [14] Meanwhile, the 802.11a has its unique advantages and was still used in certain areas. A device that supports a/b/g standards is also currently available. [08][09]

IEEE 802.11n is the newest standard of WLAN (Wireless Local Area Network) next generation. In January 2006, Hawaii, IEEE passed a draft of the 802.11n standard. However the official 802.11n universal standard is still in discussion. There are two main groups working on it, TGn Sync and WWiSE. In order to complete the standard earlier, these two groups composed a team called IPT (joint proposal team). [02]

The following section is going to introduce the capabilities of IEEE 802.11a/b/g theses three standards. Table 3-1 shows the IEEE 802.11 standard family members. [01][86]

Table 3-1 Family of IEEE 802.11

Throughput Data Rate Protocol Release Date Op. Frequency (Typ) (Max) Legacy 1997 2.4 GHz 0.9 Mbps 2 Mbps 802.11a 1999 5 GHz 23 Mbps 54 Mbps 802.11b 1999 2.4 GHz 4.3 Mbps 11 Mbps 802.11g 2003 2.4 GHz 19 Mbps 54 Mbps 802.11j 2004 4.9 - 5 GHz 23 Mbps 54 Mbps 5,15 – 5,35 indoor 802.11h 2004 23 Mbps 54 Mbps 5,47 – 5,725 indoor/outdoor 802.11n Sept, 2008 (est.) 2.4 GHz/5 GHz 74 Mbps 248 Mbps 802.11y March, 2008 (est.) 3.7 GHz 23 Mbps 54 Mbps

3.2. Capacity

As shown in the Table 3-2 [16], although the operation band of 802.11b is in 2.4 GHz, the FCC (Federal Communications Commission) is only allowing it to carry a certain level of frequency. Therefore for the U.S. the available WLAN channels are from 1 to11. This rule is also used on 802.11g. When building a WLAN, the interference is a very important factor. To avoid interference between channels in one area, the channels should not be overlapped and the interval of the frequency must be at least 25MHz. [04][05][16] That means, when selecting a channel, the number has to be five numbers apart from each other, for example, 1, 6 and 11. Thus, for 802.11b, the maximum bandwidth capacity is 3*11 = 33Mbps and for 802.11g is 3*54 = 162 Mbps. [11][14][17] - [20]

The UNII is the initial for ‘Unlicensed National Information Infrastructure’. This band is only free in certain countries, such as the U.S. and Taiwan. Table 3-3 [16] In the U.S., this band is divided into three sub-bands and then each sub-band is split up into four non-overlap channels. Therefore, the total available channels for 802.11a are twelve. Since the 5 GHz band is not commonly used, it has a lower level of interference. When using 802.11a to build a WLAN, one

11 area can have a maximum of eight non-overlap channels to use, so the overall bandwidth capacity is 8*54 = 432 Mbps. [08][09]

Table 3- 2 Capacities for 802.11a/b/g

Standard 802.11a 802.11b 802.11g Term

Band 5 GHz 2.4 GHz 2.4 GHz

Available channels 12 11 11

Bandwidth capacity 432 Mbps 33 Mbps 162 Mbps

Coverage 75 ft. 125 ft. 155 ft.

Datarate 54 Mbps Max. 11Mbps 54 Mbps

ERP-OFDM Modulation OFDM DSSS DSSS-OFDM

Table 3-3 IEEE 802.11b/g Channel Used for Different Countries

Country Name Frequency Channel Number North (GHz) Japan Europe France Spain America

1 2.412 X Y Y X X

2 2.417 X Y Y X X

3 2.422 X Y Y X X

4 2.427 X Y Y X X

5 2.432 X Y Y X X

6 2.437 X Y Y X X

7 2.442 X Y Y X X

8 2.447 X Y Y X X

9 2.452 X Y Y X X

Table 3-3 Cont.

10 2.457 X Y Y Y Y

11 2.462 X Y Y Y Y

12 2.467 X X Y Y X

13 2.472 X X Y Y X

14 2.484 Y X X X X

Table 3-4 IEEE 802.11a Channel Used for North America

Band (GHz) Frequency (GHz) Channel Number

5.180 36

5.15 – 5.25 5.200 40 UNII Lower Band 5.220 44

5.240 48

5.260 52

5.25 – 5.35 5.280 56 UNII Middle Band 5.300 60

5.320 64

5.745 149

5.725 – 5.825 5.765 153 UNII Upper Band 5.785 157

5.805 161

802.11b/g is the most widely used standard for WLAN. Because of the different operation band, 802.11a cannot cooperate with b/g standard. Also, there are some other problems, such as its high cost for equipments. The working band is also not free for all countries, and has a lower transmission distance. However it is still used currently, because it is fast, stable, and has high capacity. Also, some people think that the lower transmission distance makes the network more

13 difficult to be hacked; therefore, it is popular for some industries. To improve the competitiveness of 802.11a, researchers invent a triple frequencies chip for a/b/g. Someone who works in the office with 802.11a can then walk out to a public place and get on the Internet with 802.11b/g. The new products make certain that these three standards can be operated in one building. [7] – [9] [14]

3.3. The Physical Layer and MAC Layer 3.3.1. The Physical Layer 3.3.1.1. Introduction

The Physical layer is the first layer of the OSI (Open System Interconnection) model, figure 3-1. It is at the bottom of the model. This layer is responsible for deciding the transmission method, bandwidth, mediums, bit synchronization, modulation and so on. Most physical set up is completed in this layer. It is the first layer of the network structure. It provides a mechanical, electrical and also procedural flattop between the transmission medium. The most important mission of the physical layer is the fluency of data streams. This section is going to introduce several technologies used for 802.11 protocols and will discuss how they work in this layer.

Figure 3-1 OSI Model

3.3.1.2. PLCP and PMD

The physical layer contains two sublayers, PLCP (Physical Layer Convergence Protocol) and PMD (Physical Medium Dependent), figure 3-2. These two layers communicate with each other by SAP (Service Access Point). Different transmission methods will add different preambles and headers to their own PLCP and PMD. [04] – [07][09][21]

z PLCP

The function of PLCP is to prepare MPDU’s (MAC Protocol Data Units), also called PSDU (PLCP Service Data Unit), for communicating with the MAC layer. It transfers the incoming frames from mediums to the MAC layer. The PLCP is located between the PMD sublayer and the MAC layer so that it can help PMD transmit frames without communicating with the MAC layer in advance. For this purpose, PLCP will map the MPDUs into PMD frames before transmission. MPDU’s contain a unique preamble and header in the physical layer, which includes the information of PHY transmitters and receivers. These kind of composited frames are called PPDU (PLCP Protocol Data Unit).

z PMD

This sublayer is right under the PLCP. It controls the transmission and reception of the physical layer data units from stations via mediums. PMD is the interface of a layer and a physical medium. It connects with wireless mediums (RF, … etc.) directly. Frames are also modulated and demodulated in this sublayer.

Figure 3-2 802.11 PLCP, PMD and MAC Structure

3.3.1.3. CS/CCA

No matter what kind of network it is, there is one problem in common. That is how to improve effective transmission. For solving this problem, 802.11 applied CS/CCA technologies during transmission process. CS/CCA is the initial of Carrier Sense/Clear Channel Assessment.

CS/CCA can detect the state of a medium and report it to the transmitter. It is activated when the receiver and transmitter are on but no data streams are currently passing. It determines the channel state before transmitting. If the channel is busy, it will wait for a period of time, and then detect it again. CS/CCA also can detect whether a signal can or cannot be received by a receiver. [04][05][21][23]

3.3.1.4. IEEE 802.11b, DSSS and HR-DSSS 3.3.1.4.1. Theory and Transmission method

DSSS (Direct Sequence Spread Spectrum) was first added to the 802.11 standard in 1997. At that time the speed was only 1 Mbps and 2 Mbps. But soon, it was found that DSSS has potential to run faster. The next standard 802.11b with new DSSS came out in 1999 and ran at 5.5 Mbps and 11 Mbps. Although these four speeds are often combined together as one standard, they actually belong to two different standards.

The principle of DSSS is to transmit a signal over a wide frequency band, by spreading the RF energy across a wide band precisely. Then a receiver can get the transmitted signal by operating the correlation process.

Figure 3-3 DSSS Transmission [17]

For transmitting a signal, first a spreader has to flatten the amplitude of the narrowband radio signal and spread its RF energy to a wide band. This step includes a lot of mathematical calculations. After this process is complete, the signal will look like a RF low level noise. Then when the receivers monitor the wide frequency band and locate the noise-like signal, they can identify it as the transmitted signals/data. The received noises will be recovered by a correlator which can invert the spreading process. Since the true noise can not affect the whole band, the correlator can spread it out without damage the original signal.

When modulating the DSSS data streams, 11-chips will be added to the transmitted signals. A chip is a binary number and it is just only a part of encoding and transmission process. Chips do not carry any data. The chipped streams have another name called PN Codes (Pseudorandom

Noise Codes). The most powerful consumptions of the DSSS-PHY are to generate chipped data streams and recover data streams from chipped streams. [24] – [28][16]

“Spreading ratio” is an important parameter used to decide how many chips are needed for one bit. The improper spreading ratio will cause the waste of bandwidth. Although to increase the spreading ratio does help the ability of recovering data, it will also require higher chipping rate and larger frequency band. And there are prices to pay for increasing the chipping rate. First is the hardware, the costly high frequency RF components and second is the cost of enlarging frequency bandwidth.

The modulation method used for DSSS are DBPSK (Differential Binary Phase shift Keying) for 1 Mbps and DQPSK (Differential Quadrature Phase shift Keying) for 2 Mbps. The encoding method is Barker code. [24] – [26][16]

3.3.1.4.2. Against interference

To avoid interference and guarantee the fluency of the data streams, the two channels, according to 802.11 rules, must be at least 25 MHz away. The overlapped channels will cause more interference and damage the transmitted frames. By using DSSS transmission the interference problem have been improved to a certain extent. During transmission, because chips have been added to the data streams, they can get protection from chips. It is exactly like the function of armor. Even if the chips are damaged, the data is still safe. However, if the interference is really strong, it can still hurt the data very badly, and then nothing can be recovered. [27][28]

3.3.1.4.3. PLCP and PMD of DSSS

The DSSS-PLCP adds six parts in the DSSS-PPDU. The first two parts are PLCP preamble and the other four are headers. The structure is showed in Figure 3-4[21]. The data rate of DSSS- PLCP is 1 Mbps and 2Mbps. This preamble is also called a long preamble. The function of the preamble is to help a receiver to synchronize to the incoming frame. The header contains the information of the frame.

The DSSS-PMD is used to operate transmission and reception between stations. It is also in charge of modulating and de-modulating the PPDUs. The relationship between PLCP and PMD is like a brain and a body. PMD is response for action and PLCP gives an order. [04] – [06][16][21]

Figure 3-4 DSSS PLCP

3.3.1.4.4. HR-DSSS and CCK

The original modulation method (Barker code) used in DSSS, for 802.11, only achieved the speeds of 1 and 2 Mbps. For commercial use, this is not an acceptable amount. In the data stream, each barker word carries one or two bits depending on the data rate. To improve the speed, it is necessary to increase the capacity of each symbol. That means more complicated phase angle changes. For example, the DQPSK receiver could at least detect four cycle phase differences. More cycle phase differences will cause the smaller phase changes. It ends up with the need of a highly accurate and costly receiver.

Therefore, the next standard 802.11b had a different modulation method known as Complementary Code Keying (CCK). This method splits the data stream into code symbols composed with 8 bits. By a complex mathematic procedure, the CCK can encode four or eight bits per code word with 8-bit code symbols. Also, this procedure can help receivers to identify different codes easily, even in interference or a multipath fading environment. By using the CCK process, the speed can reach 5.5 Mbps or 11 Mbps.

The process of CCK is quite similar to the DS chipping process. Figure 3-5[17]. The difference between them is that CCK does not use the static repeating code word, as for example, Barker Code. The code word is converted from the data. These code words are used to transmit data and spread the signal. Phase angle also plays an important role here, but this time it will not cost as much as DQPSK.

Figure 3-5 CCK Modulation

HR-DSSS PHY is defined as the data rate equal to 5.5 or 11 Mbps. It also contains two parts, PLCP and PMD. PLCP is in charge of framing and PMD is response for transmission.

3.3.1.4.5. PLCP and PMD of HR-DSSS

The structure of HR-DSSS PLCP is show in Figure 3-6[17]. It is very similar to DSSS- PLCP, the only difference being in preamble. Its preamble is called short preamble. This kind of preamble could improve the efficiency of PLCP framing.

For backward compatibility, the HR-DSSS PMD can support both low (1, 2 Mbps) and high (5.5, 11 Mbps) transmission. The only difference is low speed must uses long preamble while high speed uses short one. In addition, for high speed, the framed data will be encoded by CCK modulation before transmitting. At 5.5 Mbps, each symbol has four data bits and at 11 Mbps, each symbol has eight data bits.

Figure 3-6 HR/DSSS PLCP Framing

3.3.1.5. IEEE 802.11a and OFDM 3.3.1.5.1. Background

Because there are lots of non-802.11 signals that also existed in the 2.4 GHz band, this band is often very crowded. For achieving higher speed and capacity, 802.11 Task group A (TGa) came out with a new standard. It used a 5 GHz un-licensed band and a different transmission method (OFDM). Although this standard was published in 1999, the practical hardware was not available until the end of 2001. At first, 802.11a used the UNII band and was specified for the U.S. only. Later, two similar standards were published. 802.11h is for Europe and 802.11j is for Japan. [08][30][31][43][44]

OFDM (Orthogonal Frequency Division Multiplexing) was developed in the late 1960s. It is not a new theory however, because of the wireless technology it revived again. OFDM has several features. First, it uses multiple subcarriers to do a single transmission. Second, for simplifying equalization at receiver, it transforms scattered channels into parallel narrowband subchannels. Third it is based on simple mathematics. OFDM is related with FDM (Frequency Division Multiplexing), that both divide bandwidth into divisions which is called carriers or subcarriers and use them for transmitting information. [08][17][29][33][36]

3.3.1.5.2. Principles

When transmitting data, OFDM chooses channels which overlap in frequency domain, but will not disturb with each other. Figure 3-7[17] These overlapping subcarriers are defined by mathematical calculation, therefore, they can travel individually, and this relationship is called orthogonal. With this feature, OFDM can increase capacity in the fixed bandwidth. Thus the overall performance is improved. [08][17][29]

Figure 3-7 FDM and OFDM

Although OFDM has better performance, there is still a cost. Inter-symbol interference (ISI) is a common problem when transmitting data. It mainly occurs when the delay that happens in different paths is too large and causes a later copy to shift onto a previously arrived copy. The

20 other problem is inter-carrier interference (ICI). When several subcarriers transmit in a channel and one of them shifts slightly then the interference between subcarriers would be occurred.

To avoid these problems, scientists came out with several methods. The first method was to add a “guard time” between each subcarrier. A guard time is a silent period between two subcarriers, however this method wastes bandwidth and once the delay is too large then the orthogonality would be destroyed. The other way to avoid interference is using “cyclic prefix (CP)” and “cyclic postfix (PP)”. Figure 3-8[51] CP is a guard interval between the signal N and the signal N-1. PP is an interval in signal N+1 and N. The idea of CP is to extend the time range of signal N longer, so that even if interference happened it will not affect FFT block which is the information section of a signal carried. However, sometimes the problem not only comes from the previous signal, therefore, the guard interval will have two parts “cyclic prefix” and “cyclic postfix”. Here are some other studies about OFDM and interference: [30] – [40]

Figure 3-8 CP and PP

OFDM modem is based on block-by-block structure. The transmitter first converts the signal into N subcarriers and then uses IFFT (inverse Fast Fourier Transform) to modulate them into orthogonal waveforms. The receiver reverses the process by using FFT (Fast Fourier Transform). It will remove CP first and find the FFT block. Then it starts using FFT to demodulate the signal that has received. Finally, the information is recovered. Figure 3-9[51]

OFDM has higher capacity and is more reliable than some of other methods. The mathematics of OFDM is FFT and IFFT; and they are not complicated. With FFT, the operation number in each signal is on the order of NlogN. This can be done easily by a program. That is the reason that OFDM is broadly used in wireless technology. [34][36] - [39]

Figure 3-9 OFDM Modem

3.3.1.5.3. The IEEE 802.11a’s OFDM and its PLCP and PMD

OFDM is a good transmission technology, but TGa did not use the full version. They modified their own one.

When building an OFDM network, there are some important parameters which are bandwidth, delay and bit rate. The bandwidth is fixed for 802.11a. When considering delay, it is necessary to have guard time. Usually the ratio of guard time and delay is two to four times as big. And the symbol time is five times of the guard time. Thus TGa chose 800 ns as guard time and 4 µs as symbol time. The bandwidth of an operation band is 20 MHz and the theoretical highest speed is 54 Mbps. Higher capacity could have higher throughput. However it will decrease the number of operation channels. Therefore, how to achieve a reasonable balance is an important design issue. [17][29] - [31][39]

The modulation technology for OFDM is shown in Table 3-4[17]. To reach a higher speed, TGa has not only changed the transmission method, but also used another modulation technology QAM (Quadrature Amplitude Modulation). Figure 3-10[55] For preventing the BER (Bit Error Rate) increase which caused by certain channel’s signal fading, the error correction code is added to the all sub-channels. The error correction code is called “Convolutional Code”, and this kind of OFDM sometimes is named COFDM (Coded OFDM). Here is the related research: [36][39][40] – [42]

Table 3-5 OFDM Modulation and Data Rate

Figure 3-10 Constellations Diagram

The OFDM-PLCP structure is shown in Figure 3-11[21]. The preamblle contains 12 symbols and is used to synchronize with a receiver. The duration time of an OFDM preamble is 16 µs. There are two parts of the preamble; short training sequence (the first ten symbols) and long training sequence (the rest two symbols). The short one is used to create AGC (Automatic Gain Control), timing and initial frequency offset estimation of the carrier signal. The long one is used

23 for the channel, timing and fine frequency offset estimation. The Head contains two parts, the signal and data’s service section. The head is 40 bits long (signal is 24 bits plus service 16 bits). The signal field is always transmitted with BPSK at 6 Mbps.

Figure 3-11 OFDM PLCP Structure

In the OFDM-PMD sub-layer, to maintain the speed 6, 12 and 24 Mbps is required. An operation channel (20 MHz bandwidth) is divided into 52 sub-channels (sub-carriers). Four of them are pilot carriers; the rest of the 48 sub-channels are used to transmit data. The data will be divided and modulated into sub-channels, then transmitted parallel. [17][21][29]

3.3.1.6. IEEE 802.11g and the Physical Layer 3.3.1.6.1. Background and Backward Compatibility

As 802.11b became more and more popular, people started to notice the convenience of wireless LAN. At that time, 802.11b could only provide a maximum 11Mbps. Compared to Ethernet, this was not a good performance – it was just barely acceptable. When 802.11a started to join the market, end users also wanted 802.11b to become faster. In fact, 54 Mbps really was not very fast and most of the time, in a wonderful environment, the performance could only achieve 50% of the possible speed. However WLAN has its own advantages and “wireless” has become a trend gradually. Under these conditions, 802.11g is the answer. It has the same speed with 802.11a, but it has a longer area of coverage. Moreover, it operates in the 2.4 GHz band.

According to 802.11g clause, this standard is not a “new” standard. It is an integrated standard using many existing technologies. Its physical layer is called ERP (Extended Rate PHY)

24 and contains DSSS, CCK, PBCC (Packet Binary Convolution Coding) and OFDM. It supports the speed:

ERP-DSSS: 1, 2, 5.5 and 11 Mbps

ERP-PBCC: 22 and 33Mbps

ERP-OFDM: 6, 9, 12, 18, 24, 36, 48 and 54 Mbps (6, 12 and 24 Mbps are mandatory)

For backward compatibility, these methods have some slight changes, but are pretty much the same original standard. An 802.11g stationary must have the ability to communicate with both old 802.11b stations and other 802.11 stations. For that reason, the ERP-DSSS and ERP-CCK were added.

A problem of backward compatibility is that the 802.11g can receive and decode 802.11b signals, but the converse is not true. Therefore a protection has to be added to the 802.11g standard. The first part of the protection is that when doing transmission with 802.11b stations, the Beacon frames cannot be transmitted higher than 11 Mbps. The second part is to avoid network interference between 802.11b and 802.11g. When 802.11g is transmitting data, in order to avoid disturbing 802.11b, it will send CTS (Clear To Send) frames to notify the 802.11b stations and update the NAV (Network Allocation Vector).This process is called CTS-self Protection. To make sure every station in the network can receive and process the CTS frames, they have to be sent under 802.11b protocol. The protection is controlled by the ERP information element in Beacon frames and will limit the data rate of 802.11g. [13][14][17][20][21][29]

3.3.1.6.2. ERP-OFDM and DSSS-OFDM

The ERP-OFDM is very similar to the OFDM used in 802.11a with only a slight difference. The PLCP is shown in Figure 3-12[21]. When comparing these two OFDM-PLCP, the difference is the “signal extension” field. This 6µs extra time is used to prepare and decode the 802.11a frames. 802.11a uses 16 µs SIFS (Short Inter-Frame Space) which is different form 802.11b 10 µs SIFS. 802.11g chooses 10 µs because of backward compatibility, therefore, it adds 6 µs after a frame to match timing and frame with 802.11a. Beside the PPDU structure, the transmission method is exactly the same with 802.11a.

Figure 3-12 DSSS-OFDM PSDU Format

DSSS-OFDM is a composite framing technology. Its PPDU structure is shown in Figure 3- 13[21] and Figure 3-14[21]. The preamble and head are framed by HR-DSSS mode, and the PSDU is the ERP-OFDM PPDU. The preamble has long and short formats. With this method, the protection is not required. The data rate of DSSS-OFDM preamble and head is the same with DSSS. [13][14][17][20][21][29]

Figure 3-13 DSSS-OFDM Long Preamble Structure

Figure 3-14 DSSS-OFDM Short Preamble Structure

3.3.2. MAC Layer 3.3.2.1. Introduction

In the human world, laws are the tools to avoid arguments and accidents and it is the same with networks. Wireless transmission has many problems such as interference, fading, collision, and security. In the MAC (Media Access Control) layer, the 802.11 working group, established a variety of rules and functions to improve system performance and solve these problems.

3.3.2.2. DCF (Distributed Coordination Function) and PCF (Point Coordination Function)

DCF is the core of CSMA/CA and most of the transmissions using 802.11 standards are based on it. With DCF, there is no central control. Stations have to compete for the use of channels. Any transmissions using CSMA/CA (Carrier Sense Multiple Access/Collision Avoidance) belongs to DCF type transmission. The other mode, PCF, is a central control type. With PCF, a base station controls all transmission orders; therefore there is no competition and collisions in this mode. The base station will poll other stations to see if they have frames to send. The polling frequency and order are decided by a polling list, which is not equaled. Any station will be added to the list after connecting to the base station. [17][55]

3.3.2.3. Hidden node and CSMA/CA

When a station is communicating with another station, there is an existing risk that the packets may collide. It is due to the hidden node problem. Each station has limited coverage, so stations may not be aware of the other nearby stations. Thus if there are more than two stations that try to transmit a signal to the same station at the same time, and then the collision will happen. In Figure 3-15, station A and C are hidden nodes to each other. If station A is sending a signal, station C cannot be aware of it. Since most stations are half-duplex (cannot transmit and listen at the same time), the collision is difficult to be detected. In Figure 2-17, it shows that only station B knows collisions happened. [17][55]

Figure 3-15 The Hidden Node Problem

To deal with this problem, 802.11 WG uses a different method from Ethernet. With wired transmission they use CSMA/CD (Carrier Sense Multiple Access/Collision Detection), but since the wireless transmission has very limited resource and is not as reliable as Ethernet, 802.11 WG turns to use collision avoidance (CSMA/CA).

There are two technologies used in CSMA/CA, physical channel sensing and virtual channel sensing. [17][55]

z Physical channel sensing

When a station wants to send signals, for example, it senses the channel first. If the channel is idle, it just sends. During the transmission, it will not sense the channel at the time, but keeps sending signals. However if the channel is busy, it will wait until the channel is idle and then start the transmission. Once a collision occurs, the stations will wait for certain time, by using the Ethernet binary exponential back off algorithm, and then try again later. [17][55]

z Virtual channel sensing

he other method, virtual channel sensing, is shown in figure 3-16 [55] and figure 3-17[55]. When station A wants to send signals to station B, it will send a RTS (Request to Send) frame to the station to occupy a channel first. If the channel is idle, station B will send CTS (Clear to Send) frame back. When station A receives CTS, it will start to transmit a signal. When the process is initiated, an important “timer” is including in RTS and CTS frame which called NAV (Network Allocation Vector). The information that is included in NAV is how long the channel will be occupied. One operation contains RTS, CTS, data frames and ACK (Acknowledgement). NAV will set how long it will take including the last ACK frame. Therefore, since station D is within the range of station A, it will receive RTS and then stop completely from occupying the channel until NAV counter becomes zero. It is the same with station D. Station C will receive CTS, so it will also wait until the operation is completed. However, before the ACK is received by station A, the NAV is expired. Then the operation has to run again. [17][55]

Figure 3-16 CDMA/CA

Figure 3-17 Virtual Channel Sensing

3.3.2.4. Fragmentation

In most occasions, the frames from the upper layer maybe longer than the fragmentation threshold. In this situation, the frames have to be fragmented before transmitting. The advantages for using fragmentation are that it can improve throughput and reliability. By fragmenting frames, the interference only destroys small pieces of frames, but not the entire frames. Thus the overall effective transmission is improved. Also, only the destroyed pieces need to be retransmitted, therefore, the reliability is also improved.

The transmission process of fragmented frames is called fragmentation burst. In the clause, there are no rules for fragmentation threshold. It depends on network designers. During the transmission, each fragment has the same frame sequence number and an increasing fragment number for recombination. To keep occupying the channel for completing fragmentation burst, each fragment will reset the NAV for next fragment including ACK. Figure 3-18[55]. [17][55]

Figure 3-18 A Fragment Burst

3.3.2.5. Interframe spacing

Interframe spacing is a very important protocol for coordinating medium access. There are four different intervals which are shown in figure 3-19[55]. To avoid collision, stations wait a certain period of time until channels are idle before transmission. These different intervals decide the priority for different types of transmission. High priority transmissions wait a shorter amount of time, so it can occupy the channel first. [17][55]

Figure 3-19 Frame Interval for IEEE 802.11

z SIFS (Short InterFrame Spacing)

This is the shortest interval and mainly used for high priority transmission; including a receiver sends CTS back to a transmitter, a receiver sends an ACK for a fragment or a full frame and a transmitter of a fragmentation burst send the next fragment without sending RTS again.

z PIFS (PCF InterFrame Spacing)

This interval is used for competition-free transmission by PCF. If the SIFS interval passed and no stations occupy the channel, after PIFS, the base station will send a poll frame or a beacon frame. Any stations which have data or fragment sequence waiting for transmitting can start to send after PIFS without disturbance or competition.

z DIFS (DCF InterFrame Spacing)

After DIFS, every station can start to compete for earning the channel usage and then the winner takes all. The all transmission rules and Ethernet binary exponential back off algorithm are applied.

z EIFS (Extended InterFrame Spacing)

This is only used when a mistake has happened during a transmission.

3.3.2.6. Power saving

To extend the battery’s life of the MS (Mobile Station), the 802.11 MAC layer provides a power management protocol. A base station can direct a MS to go into the sleep mode and buffer frames for it. Later the MS could be awakened by a base station which sends a beacon frame or a

31 user. The awakened MS will send a PS-Poll frame to the base station to get the buffered frames. The base station can choose either immediate response or deferred response to the MS. [17][55]

The immediate response means the BS (Base Station) will send the buffered frames to MS after SIFS interval. If the BS chooses a deferred response, it will send an ACK frame back to MS first, and then transmit data frames later. After sending the PS-Poll frame, the MS must stays awake until the whole process is over. The BS notifies the MS for buffered frames by sending a beacon frame. The buffered frames may be fragmented for transmission. [17][55]

3.3.2.7. Security

This is a very important part for the whole MAC layer design. So far the most widely used security procedure for wireless communication is WEP (Wired Equivalent Privacy). WEP uses RC4 cipher to encrypt data. The RC4 cipher is a kind of symmetric stream cipher; it generates a keystream and then uses the XOR algorithm to mix with data to produce the ciphertext stream. The receiver will use the same XOR algorithm to recover original data. To encrypt the data, the secret key has to be chosen first and then extended to the same size of data by PRNG (Pseudorandom Number Generator). This extended secret key is called keystream. For recovering data, both transmitters and receivers must have the same secret key and PRNG, and how to distribute the secret key is an important issue; sometimes it may be preloaded by system designers or manufacturers. The other issue is “key to keystream expansion” of RC4 stream cipher, because the safety is dependent upon on how random it is. [17][55]

The communication security has three major properties: confidentiality, integrity and authentication. Confidentiality is needed to protect data from being stolen by unauthorized people. Integrity is used to make sure the data has not been changed during transmission. This part is dependent on CRC code. Finally, authentication is the foundation of all security procedures. For transmitting data, the users must be trusted and the source must be reliable. Otherwise, authorization and access control will not be allowed. [17][55]

RC4 shared secret key is composed by 40-bit shared secret and 24-bit IV (Initialization Vector) usually called 64-bit WEP and the other one is 128-bit WEP. When framing a frame, WEP will generate an ICV (Integrity Check Value) which is a hash value to combine with payload as an original data. This ICV can use to protect data from unauthorized changes. After entering the original payload, secret key and IV, WEP can generate an encrypted frame for transmission in either secure or not secure network. Figure 3-20[17] shows the frame structure and operation of WEP. Figure 3-21[55] shows the encryption process. [17][55]

Figure 3-20 WEP Frame and Operation

Figure 3-21 The WEP Encryption Process

However there is no perfect encryption in the world. There are some problems that exist in WEP; key management, reusing the keystream, and IV. WEP has been thought of as a very secure way to protect wireless communication. However, a study from University of California, Berkeley called ISAAC (Internet Security Applications, Authentication and Cryptography) showed the defects of WEP design [56]. Also a method for breaking WEP has already been published [57].

The reason that WEP was safe was because the secret key was not easily broken. However in some cases, the network designers use the same shared key for all users. That means that people can read each other’s packets. That is not good for security. If it is a small network, trying to use the different keys for each user should not be a big problem, but if it is a big network, for security reasons, if a secret key changes once in a period of time, it is a huge job. How to distribute a secret key and how to assure it is not repeated are very important issues at this point. Moreover, IV is another potential risk. The WEP protocol recommends that the IV value should be different for every packet to avoid the keystream attack. If IV is not a random value, the encryption will be broken easily. Unfortunately some wireless network cards set the initial IV as 0 and increase the number by one for each packet sent. With this kind of IVs, to break the encryption is way too easy. Unless the secret key is changed often or assuring the use of random and different IV for each packet, the risk will remain high. If a hacker collects enough packets with repeated IV from the same user, the cipher will be broken. [17][55]

3.4. The next generation standard 3.4.1. IEEE 802.11n

IEEE 802.11n is the standard of WLAN for next generation which also called Wi-Fi (Wireless fidelity). This standard was initiated in 2004 by TGn. The goal of this standard was to improve the net transmission speed up to 100 Mbps. In the beginning, there were six proposals that had been posted. After many discussions, there were only two proposals left and they were from TGnSync and WWiSE. These two groups have their own support from chip makers. TGnSync has Athero, Agere, Marvell and Intel, while WWiSE has Airgo, Broadcom, Conexant and Texas Instruments. The first draft was approved in March 2006 and draft 2.0 was approved in March 2007. Since the demands from the market are becoming higher, Wi-Fi Alliance decided to start to certifying IEEE 802.11n products based on the draft 2.0 in summer 2007. Now, these pre-n products are already in stores, such as routers and network cards for desktops or notebooks. The official approval of IEEE 802.11n standard had been delayed many times because TGnSync and WWiSE kept fighting for their own proposal to become the official standard. However, they finally agreed to propose a merged proposal to speed up the birth of official IEEE 802.11n standard. Therefore, the IEEE 802.11n standard may be finally approved in early 2008. [17] [65] [66]

Actually, the ideal of these two groups have may be different, but the technologies are the same. TGnSync emphasizes on improving the peak data rate and WWiSE wants to ameliorate the MAC-layer. Both of them can reach the goal, or even much better than that. The common points are they both use MIMO-OFDM as the transmission function and support the bandwidth for both 20 MHz and 40 MHz. IEEE 802.11n PHY-layer is extended from IEEE 802.11a PHY- layer. So it is also an OFDM-based PHY-layer and just introduced MIMO technology to increase the speed. [17] [61]-[66]

3.4.2. MIMO (Multiple-Input/Multiple-Output)

Usually, the IEEE 802.11 air-interface only transmits data with single antenna. Although some systems use two antennas, systems that only use the one have the best performance. Therefore, no matter how many antennas that systems has, there is only one used to transmit and receive data, and there is only one input chain and one output chain.

The basic operation of MIMO is to distribute the RF chain to every system antenna and each RF chain can do simultaneous transmission and reception. This can enormously increase throughput. Moreover simultaneous receiver processing can solve multipath interference and improve the quality of the received signals. A frame can be split, multiplexed and then transmitted by more than one spatial stream. The antenna configuration of MIMO usually express as “YxZ “format. For example, in both TGnSync and WWiSE proposals require “2x2” operation, which means there are two transmit chains, two receiving chains and two spatial streams. That is mandatory mode in both proposals and there is optional mode included. [17]

When setting up hardware, the BS and end users may have different numbers of antenna. For example, a BS has three and the end user has two. Usually BS has more antennas, because of saving power and cost for the end users. In this situation, 2x3 is for uplink and 3x2 is for downlink. Two spatial streams are distributed to three antennas. One spatial stream is transmitted by multiple antennas; this is called STBC (Space-Time Block Coding). This method is also used in IEEE 802.16 PHY-layer. [17] [61]-[66]

3.5. Limitation

As the Internet goes into the wireless stage, internet access seems become more and more convenient. Wired Internet access, such as ADSL, can provide at least 100 Mbps data rate, and Wi-Fi, with the g standard, can achieve 54 Mbps, and it depends on the air-interface conditions however, if Wi-Fi wants to share the ADSL market, it still has a long way to go. Wi-Fi is strict to the operation environment and sensitive to channel fading, so reliability is a big issue. Since Wi- Fi is designed for small area not BWA, such as home or office, VoIP (Voice over IP), video service and a large file download are heavy duty for it. If a business building wants to install a wireless network for the whole building, every floor may need an individual AP (access point) or even more than one, depending on the location structure. This is because the signals suffer the short transmission range, channel fading and interference, therefore the installation cost and complexity will be high.

The other problem of Wi-Fi is the market share. This is caused by fundamental infrastructure, completion rate users’ behavior patterns and technologies maturity. What is Wi- Fi’s market positioning? And who is the major customer group? The answers of these two questions are the key to the future of Wi-Fi. If Wi-Fi wants to take VoIP, for example, as the kill

35 application, its biggest opponent will be 3G. Or if Wi-Fi wants to take broad area wireless network as the goal, the WiMax is blocking its way. Compared to 3G, Wi-Fi is faster, but the signal is not as stable. And compared to WiMax, Wi-Fi is slow, has small-coverage and is unreliable. In this difficult situation, Wi-Fi should not keep thinking its killer application or try to occupy market share of other wireless technologies. The future of Wi-Fi should be cooperation not competition.

Since Intel integrates Wi-Fi with laptop computers, it speeds up the demand of wireless access. The laptop has come into the main stream in the IT industry. People can carry their PC with them, but they cannot carry the ADSL. Therefore, it is the reason why a wireless network is desirable. Generally, most people believe that businessmen are the one who need wireless access everywhere. However, the major Wi-Fi users or the potential users should be home workers and students. Observing the behavior patterns, it is easy to find out that people want to get rid of annoying wires and use the Internet access in different rooms at home, and students need to access the Internet for information, online games or entertainment in different indoor hotspots at any time. Another important fact from the behavior patterns is there is no killer application for Wi-Fi. People access the Internet for enjoying ubiquitous service not just for a single purpose. Thus, with the new coming standard IEEE 802.11n, Wi-Fi should focus on how to get involved to meet the demands of home users and students. That means to provide ubiquitous access.

Therefore, the next problem is if the ISP (Internet Service Provider) wants to build the public Wi-Fi network. Is it possible and what are the issues behind it? This topic will be discussed in chapter five.

CHAPTER FOUR

4. IEEE 802.16, Wireless MAN

4.1. The background of IEEE 802.16

The IEEE 802.16 standard is the second generation of BWA (Broadband Wireless Access). The standard group was formed in 1998 and its purpose was to develop an air- interface standard for BWA. At the beginning of the project, the group focused on a LOS-based (Line of Sight) point-to-multipoint wireless broadband system operated in the 10 GHz – 66 GHz band. The complete standard was finished in December 2001. The evolution of IEEE 802.16 can be split into four stages: 1. Narrowband wireless local-loop systems, 2. First-generation line-of-sight broadband systems, 3. Second-generation non-line-of-sight broadband systems, and 4. Standards-based broadband wireless systems. [50][60]

z Narrowband wireless local-loop systems

The First system is related to the wireless voice telephony. The WLL systems (Wireless Local-Loop) were successful in many developing countries such as China, India and Brazil. There are two major technologies used in these WLL systems digital-enhanced cordless telephony (DECT) and code division multiple access (CDMA). To stay competitive, WLL systems started to join the Internet service market in 1993. In February 1997 AT&T developed a wireless access system for 1900 MHz PCS (Personal Communications Service) and ended the service in December 2001 due to high costs and poor take-rate. During the same time, some small companies focused on wireless internet service. These WISP (Wireless Internet Service Provider) companies set up the system in license-exempt bands, 900 MHz and 2.4 GHz and needed customers’ permission to install antennas either on the rooftop or top of the building. At this time the range, capacity, and speed were limited. [50][87] – [90]

z First-generation line-of-sight broadband systems

Since wired internet service can provide higher and more stable service, wireless systems needed to evolve to be competitive. There are two major systems called local multipoint distribution systems (LMDS) and multichannel multipoint distribution services (MMDS). LMDS mainly supported SOHO (Small Office, Home Office), business centers, and small corporations. This system only had short success in the late 1990s. MMDS was once used to provide wireless cable broadcast video service in rural areas where no cable TV service was available. When satellite TV came out, the wireless cable business crashed. The operators sought an alternative way to use this band (2.5 GHz). In September 1998 FCC relaxes rules for the MMDS band to allow two-way communication. After these changes in regulations some companies such as MIC WorldCom and Sprint, bought licenses to use the MMDS spectrum and began to develop high

37 speed fixed wireless broadband service for this band. These first-generation broadband systems used towers which were several hundred feet tall and had LOS coverage up to 35 miles with high power transmitters. The users had to install outdoor antennas high enough to receive a signal and pointed toward the tower for the clear LOS path. This LOS equipment was soon considered an impediment. And because towers were difficult to set up, the service was quite limited. [50][91][92]

z Second-generation non-line-of-sight broadband systems

The second-generation broadband systems had potential to solve the LOS problem and had the ability to provide more capacity and higher speed. This was because they used cellular architecture and advanced-signal processing technology. With these methods, the link and system performance was improved under multipath conditions. These new second-generation systems also perform well in NOLS (Non-Line of Sight) environment by using OFDM, CDMA and multiantenna processing. Some of them can even be operated without setting an antenna outside. [50]

z Standards-based broadband wireless systems.

IEEE 802.16 standard was approved officially on December 2001 and the formal name was wireless metropolitan area network (WMAN). This first standard only focused on the10 GHz – 66 GHz band and can only do LOS transmission. In PHY-layer, it used single-carrier modulation and in the MAC-layer, it had time division multiplexing (TDM) structure which supported FDD (Frequency Division Duplexing) and TDD (Time Division Duplexing). [50]

In order to maintain a competitive advantage, the IEEE 802.16 standard group developed a series of extended standards to improve WMAN performance. The IEEE 802.16 family members include IEEE 802.16, IEEE 802.16a, IEEE 802.16c, IEEE 802.16-2004, IEEE 802.16e-2005, IEEE 802.16f, IEEE 802.16g and IEEE 802.16h. [60]

z WiMax Forum

This is a Non-Profit Organization formed in April 2001. The initial members included Intel, Fujistu, and Nokia among others. The major mission of this forum is to assure the unity and compatibility of BWA products produced by any manufacturer. In the beginning, the certification was only between IEEE 802.16 standard and ETSI HiperMAN standard (European Telecommunications Standards Institute, High Performance Metropolitan Area Network). As the technologies of the 802.16 become more mature, many industries started to pay attention to it. Therefore, more companies around the world joined and established several working groups to promote the standard. Because of this forum, WiMax (World Interoperability for Microwave Access) became the second name of IEEE 802.16. [59]

The working groups of WiMax Forum are CWG (Certification Working Group), TWG (Technical Working Group), RWG (Regulatory Working Group), MWG (Marketing Working Group), SPWG (Service Provider Working Group), NWG (Network Working Group) and AWG (Application Working Group) [59]

4.2. Capacity of the IEEE 802.16 family 4.2.1. IEEE 802.16

This initial standard is fixed air-interface BWA. Because it operated in a high frequency band, it can only do LOS transmission. In addition, since IEEE established this unified clause, manufacturers can produce compatible equipment by following the standard. This was a big step for realizing WMAN. [45][50] [60][93][96]

4.2.2. IEEE 802.16a

This standard was approved in January 2003. The NLOS transmission was added. The IEEE 802.16 TGa extended the operation band to 2 – 11 GHz licensed and licensed-exempt band. This frequency had a longer wavelength, thus performs NLOS transmission very well. The maximum coverage was 30 miles and minimum was 6 miles. The other key technology called OFDM was also added in PHY-layer at this time. OFDM proved (in IEEE 802.11) that it was the solution for the multipath issue. These two were the major changes in PHY-layer. In the MAC-layer, the IEEE 802.16a provided QoS system the ability to support voice and video real time service. [46][50] [60]

4.2.3. IEEE 802.16c

This standard is just an improved standard of the initial IEEE 802.16 standard. This one established more detailed rules of how to run a system in 10 – 66 GHz band. [47][50] [60]

4.2.4. IEEE 802.16-2004

This standard was the first practical standard of the IEEE 802.16 family. It is also known as fixed-WiMax. It integrated the previous standards and re-edited PHY and MAC-layer contents to improve the system performance and compatibility. Therefore it can support various business uses. IEEE 802.16-2004 was a fixed air-interface of BWA, and it supported 2 – 11 GHz licensed and licensed-exempt band and 10 – 66 GHz band. That meant it can do both LOS and NLOS

39 transmission. The complete edition of IEEE 802.16-2004 was approved in December 2004. [48][50] [60]

4.2.5. IEEE 802.16e-2005

In 2003, the 802.16 standard group started another mission. The TGe began to combine mobility and BWA together. The result was IEEE 802.16e and it was formally published in February 2006. This standard introduces scalable OFDMA in PHY-layer and modifies the MAC- layer for supporting high speed mobility. The goal is to support end users moving in cars and giving them the ability to access the BWA without any problems. It has backward compatibility with IEEE 802.16-2004. Because of its mobility, it is also referred to as mobile-WiMax. [49][50] [60]

4.2.6. IEEE 802.16f, IEEE 802.16g and IEEE 802.16h

These standards are still under discussion. So far only drafts are available. [50][60] Table 4- 1[50] is the basic information of IEEE 802.16 family.

Table 4-1 The Basic Data of IEEE 802.16

4.3. The Physical Layer and MAC Layer

In the following section, the IEEE 802.16-2004 and IEEE 802.16e-2005 will be discussed. IEEE 802.16e-2005 inherits the fixed-BWA ability from IEEE 802.16-2004 and re-edits the PHY and MAC layer to combine fixed-BWA with mobility. According to the WiMax forum, there are four usage scenarios: nomadic, portable, simple mobility and full mobility. The applications of IEEE 802.16-2004 are to provide an air-interface for fixed and nomadic transmission, while IEEE 802.16e-2005 is to provide an air-interface for portable and mobile transmission. [50][59][94] – [96]

z Nomadic

The user can use his/her devices with a local fixed service and reconnect to a different fixed service in another place.

z Portable

Provide service for 3G-phones, PDA or notebook.

z Simple mobility

Allowed the end users move in the speed up to 60 km/hr with less than 1 second interruptions.

z Full mobility

Provide up to 120 km/hr mobility service with less than 50 ms latency and less than 1% packet loss.

4.3.1. IEEE 802.16-2004 4.3.1.1. Physical Layer

There are several key technologies used in IEEE 802.16-2004 (fixed-WiMax). The fixed- WiMax is based on OFDM-PHY. It has a strong ability to protect against multipath fading and support multipath and NLOS transmission. The blueprint of the operation theory is derived from the 802.11a. The 802.16 WG changed some parameters to fit its own demands. The figure 4-1 shows the fundamental structure of IEEE 802.16 bottom layers.

Figure 4-1 The Layer Structure of IEEE802.16

The other key technology used in WiMax is multiple-antenna techniques. This technology is also referred to as MIMO. Through OFDM, WiMax can achieve frequency diversity and by multiple –antenna techniques, WiMax can have spatial diversity. The term spatial diversity means using two or more antennas at the receiver and/or transmitter to do transmission. During the procedure the spatial multiplexing and the STBC is required. The idea of spatial multiplexing, figure 4-2[50], is to create multiple parallel channels to carry one data stream. The advantages of MIMO are that it increases system reliability, data rate, system capacity, and coverage, and decreases the required transmit power. The theory of MIMO was introduced in section 3.4.2. [50][95] – [101]

Where y = H．X + n

Figure 4-2 A Spatial Multiplexing MIMO System

4.3.1.1.1. Four transmission mode: SC, SCa, OFDM and OFDMA

In IEEE 802.16-2004 PHY-layer, SC (Signal Layer) is responsible for the 10 – 66 GHz band and SCa is used for the 2 – 11GHz band. The core of this standard is the OFDM/OFDMA modulation and used in the 2 – 10 GHz band. This method has many advvantages and has been widely used for many wireless technologies. [50]

4.3.1.1.2. AMC (Adaptive Modulation and Coding)

With AMC, WiMax can support various kinds of modulation and forward error correction (FEC) coding. This method allows modulation schemes to be changed for each person based on channel conditions. AMC is very effective on optimizing overall channel capacity in a time- varying channel. Therefore users can be provided with the possible highest data rate which is allowed by the ISP and link conditions. The theory is that when the channel conditions are good; then it transmits as high data rate as possible. When the channel conditions are poor, then it transmits at a lower data rate to avoid excessive packets becoming lost. Lower date rates and higher date rates can be achieved by using different modulation, such as QPSK and low-rate error-correction codes for low speed, and 64-QAM and less robust error-correction for high speed. [50]

In both IEEE 802.16-2006 and IEEE 802.16e-2005 PHY, the RS-CC (Reed Solomon – Convolution Code) is the required coding mode and block turbo codes (BTC), convolution turbo

44 codes (CTC) and low density parity check (LDPC) codes are optional modes. Since the IEEE 802.16 clause did not set up the bandwidth and modulation method, different configurations will end up with different speeds. Thus with AMC, the system can have the maximum usage of the channel. Table 4-2[50] shows the modulation and coding supported in WiMax. [50]

Table 4-2 Modulation and Coding Supported in WiMax

4.3.1.1.3. Channel-quality Measurement

The channel-quality measurement scheme is the fundamental of downlink and uplink power control processes and modulation and code rate adaptation. A MS is required to send the channel quality feedback (CQI) including RSSI (Received Signal Strength Indicator) and SINR (Signal- to-Interference-plus-Noise Ratio) to BS for evaluating channel conditions. With this information, BS can: [50]

z Change modulation and/or coding rate for the transmission z Change the power level of the associated DL or UL transmissions.

4.3.1.1.4. OFDM in WiMax

In fixed-WiMax, the FFT size is only 256. There are 192 subcarriers used for carrying data; 8 for pilot subcarriers and the rest are guard subcarriers. Because the FFT size is fixed, the subcarrier interval is related to the channel bandwidth. With larger bandwidth, the subcarrier interval increases, but the symbol time decreases. That means a larger piece of frames needs to be allocated properly to overcome delay spread. The table 4-3[50] shows the OFDM parameters used for WiMAX. The OFDM-PHY of WiMax has a wide range of guard times, therefore the network designers can make fitting trade-offs between spectral efficiency and delay spread robustness. When using a 25 percent guard time, it can deal with the maximum delay spread

45 robustness which is up to 16µs in a 3.5 MHz channel and 8µs in a 7 MHz channel. [50][102] – [106]

Table 4-3 OFDM Parameters Used for WiMAX

4.3.1.2. MAC Layer 4.3.1.2.1. Three sublayers: CS, CPS and SS

The MAC layer is an interface between the higher layer and the PHY. It also processes packets for the upper layer called MSDUs (MAC Service Data Units) and maps them into MPDUs for transmitting in PHY.The IEEE 802.16 standard splits the MAC-layer into three sublayers, figure 2-22, convergence sublayer (CS), common part sublayer (CPS) and security sublayer (SS). CS is an interface between MAC-layer and higher layer protocol such as ATM, IP and so on. CPS provides MAC management, ARQ (Automatic Repeat Quest) scheduling and assembly of MAC PDUs. SS is in charge of encryption. [50][107]

4.3.1.2.2. AAS (Advanced Antenna Systems)

Multiantena is a good way to improve system performance. However it includes some problems such as transmit diversity, beamforming, and spatial multiplexing.. The AAS is a solution for them. [50][113][114]

z Transmit diversity

Transmit diversity requires at least two transmitting antennas and one receiving antenna. For this type of transmission, WiMax defined STBC scheme, for example the 2x1 antenna system with Alamouti codes. The advantage of it is the same as with MIMO. The MS or end users will not have a complex set up and the cost is lower.

z Beamforming

Beamforming has great ability to quard against interference and improve the coverage range, capacity, reliability, and received SINR. The idea is to use multiple antennas to transmit the same signal in the direction of the receiver. To operate using this method, the signal must be weighted properly for each antenna and the transmitter has to have accurate knowledge of the channel. WiMax supports beamforming for both uplink and downlink.

z Spatial multiplexing

Spatial multiplexing means using multiple antennas to transmit multiple independent streams. When both receiver and transmitter have more than one antenna, by performing the STBC scheme, the streams can be split. The purpose of this method is to improve the data rate or capacity of the system. The increasing ratio is linear with the number of antennas. For example, a 2x2 MIMO system will double the capacity. If the MS only has one antenna, the spatial multiplexing is still supported. However, spatial multiplexing only works under good SINR conditions.

4.3.1.2.3. QoS

QoS (Quality of Service) is a basic function of the WiMax MAC-layer. By using connection-oriented MAC structure, the QoS control is robust. All the uplink and downlink connections are controlled by the BS. Before transmission, a BS and a MS set up a unidirectional logical link called “connection”. This link is between two MAC layers and every connection is verified by a 16-bit connection identifier (CID). [50]

To ensure the QoS, WiMax defines a scheme named “service flow”. It is a one way flow of packets that includes a special set of QoS parameters and is verified by a 32-bit service flow identifier (SFID). These parameters are used to support different kind of QoS of different tasks.

There are varies of parameters. [50] The SFID is issued and mapped to CIDs by the BS. There are four types of scheduling services for providing the best performance of data transfer of various applications. [50] Table 4-4[50][108] – [112]

Table 4-4 Service Flows in WiMax

z Unsolicited grant services (UGS)

UGS is designed to support real-time fixed-size packets periodic transmission at constant bit rate (CBR), for example T1.

z Real-time polling services (rtPS)

This scheme is designed to support real-time unfixed-size packets periodic transmission. An example is MPEG video. This service allows BS to provide unicast polling opportunities for the MS to request bandwidth.

z Non-real-time polling service (nrtPS)

This one supports delay-tolerant data streams with unfixed-size packets, like FTP. The nrtPS and rtPs are similar. The difference is that a MS can use contention-baased polling or unicast polling opportunities in the uplink to request bandwidth in nrtPS service.

z Best-effort service (BE)

This scheme is used only for service that does not require strict QoS, such as web browsing. The MS uses merely contention-based polling opportunities to request bandwidth and sends data whenever channels are available.

4.3.1.2.4. Security 4.3.1.2.4.1. Overview

The security of IEEE 802.16 is called the privacy sublayer at the bottom of the MAC layer. It is mainly used to provide access control and confidentiality of the data link. The configuration of the IEEE 802.16 security has five components and will be introduced in the following. [50][52][115] – [119]

z Security Associations (SA)

This component is mainly concerned about connection. IEEE 802.16 has two types of SA; data SA and authorization SA. Only data SA has a clear definition. The data SA is used to protect transmit connections between SS (Subscriber Station) and BS (Base Station).

The authorization SA is a state which is shared between two particular SS and BS. The BS uses authorization SAs to configure data SAs on the SS.[52]

For securing a transmit connection, a SS first uses a “create-connection” request to initiate a data SA. The standard will let several connection IPs share a SA to support multicast. On network entry, the standard automatically establishes a SA for the secondary management channel. Thus, a SS may have two or three SAs, one for the secondary management channel, others for uplink and downlink connections. Each multicast group requires a SA to share with group members.

z X.509 certificate profile

This is used to identify communication parities and 802.16 does not define its extensions. This standard defines two types: manufacturer certificates and SS certificates. There are no certificates of BS. A manufacturer certificate is used to identify the manufacturer of an IEEE 802.16 device. It can be self-signed or issued by a third party. An SS certificate can identify a particular SS and also its MAC address. The SS certificates are created and signed by manufacturers.

The BS uses the manufacturer certificate’s public key to confirm the SS certificate and also verify the device. The SS must take care of the private key corresponding to its public key to prevent intrusion from attackers.

z PKM (Privacy and Key Management) authorization

This protocol is used to send an authorization token to an authorized SS. And it contains three communication steps between SS and BS. Figure 4-3[52] The first two steps are sent by SS for verifying itself with BS and the third one is used for response from BS to SS.

Figure 4-3 PKM Authorization Process and Parameters

z Privacy and Key Management

PKM is used to establish data SAs between SS and BS. This protocol can have two or three steps of message exchange between SS and BS. Figure 4-4[52] The first step is optional. It depends on whether that BS requests rekeying or not. The second one, SS uses it to initiate the protocol and BS reply with the third one.

Figure 4-4 PKM Protocol Messages Exchange Process and Parameters

z Encryption

The DES-CBC (Data Encryption Standard-Cipher Block Chaining) encryption enciphers a plaintext MPDU, but not the MPDU GMH or CRC. Figure 4-5[52]

Figure 4-5 Encryption Frame Structure and Process

4.3.1.2.4.2. Overall Analysis

Learning from the experience from IEEE 802.11, the IEEE 802.16 working group faces security threats both in the MAC and PHY levels. 802.11 only has protections in the MAC level cannot present against attacks from PHY-Level. 802.16 progression has many good functions, however it also brings with it an increase of threats to users

The initial standard IEEE 802.16-2001 requires an attacker to put real equipment between SS and BS and also operate at a frequency of 10 to 66 GHz. The IEEE 802.16e adds mobility to the standard; however it creates a security risk. Now, an attacker’s physical position is not really constrained and the management messages are much weaker than IEEE 802.11.

With radio transmission, there is a great risk. Anyone who has a proper well located receiver can intercept the channel. So the designers have to establish a safety mechanism. The other threat is that anybody who has a correctly configured radio transmitter can access the wireless network. Through this weakness, an attacker can make up or modify frames from authorized organizations. Therefore there must be a procedure to protect the transmitting data. Furthermore, the interference and distance may also give attackers a chance to intrude the system or two systems which cannot connect to each other directly. Thus, how to protect a system and examine the received data are also very important to designers.

4.3.2. IEEE 802.16e-2005, the newest standard 4.3.2.1. OFDMA-PHY 4.3.2.1.1. Background

OFDMA (Orthogonal Frequency Division Multiple Access) is the main difference of 802.16e from other 802.16 standards. Most of new improvements and design are based on it. This method is deeply related with FDMA (Frequency Division Multiple Access). By using FFT/IFFT, FDMA can divide and combine the baseband in an orthogonal way. In OFDMA-PHY, the subchannels are a minimum frequency resource-unit assigned by a base station. Thus different subchannels can be allocated to different users as a multiple-access mechanism. The IEEE TGe introduces scalability to OFDMA which can help provide best performance in a channel with bandwidth between 1.25MHz to 20MHz. for both fixed and mobile service. Also it can lower the cost of system. [50][53]

The FFT size in OFDMA-PHY is scalable from 128 to 2048. When the bandwidth increases, the FFT size is increased too, but the subcarrier interval is fixed as 10.94 kHz. This also keeps the OFDM symbol duration fixed. The 10.94 kHz interval is the best choice for balancing between the delay-spread and Doppler spread requirements for operating in fixed and mobile environments. This interval allows the delay-spread up to 20µs and mobility speed up to 125 kmph in 3.5 GHz band. The FFT size 128, 512, 1024 and 2068 are used when the channel bandwidth are 1.25, 5, 10 and 20 MHz, respectively. The 256 bits OFDM is included in the OFDMA-PHY, so the mobile-WiMax is backward compatible with fixed-WiMax, when the FFT size is 256. [50][53][106][120] – [123]

Table 4-3 is the recommended scalability parameters for system.

4.3.2.1.2. Frame structure

OFDMA supports many different frame sizes to fit the need of various applications and models. There are three types of OFDMA subcarriers: [53][102][103][124] – [126]

1. Data subcarriers for data transmission.

2. Pilot subcarriers for various estimation and synchronization purposes.

3. Null subcarriers for no transmission at all, used for guard bands and DC carriers.

In different subcarrier allocation modes, the pilot allocation is performed differently. For DL (Downlink) Partially Used Subchannelization (PUSC) and all UL (Uplink) modes, the set of subcarriers of data and pilot is first sliced into subchannels, and then the pilot subcarriers are allocated by each subchannel. For DL Fully Used Subchannelization (FUSC), the pilot

53 subcarriers will be assigned first and then the rest of subcarriers are partitioned into data subchannels. FUSC has one set for pilot subcarriers, but each subchannel in PUSC has its own set for pilot subcarriers.

Figure 4-6[53] is the frame structure of OFDMA-TDD. This kind of frame is sliced into DL and UL two subframes. The DL and UL subframes are separated by Transmit/Receive Gap and Receive/Transmit Gap (TRG and RTG).The DL subframe starts with FCH (Frame Control Header) and then a DL-MAP and a UL-MAP. In DL-MAP BS will tell MS which subcarriers are going to assign for it. The UL-MAP includes which subcarriers that MS can use to transmit on.

Figure 4-6 OFDMA Frame Structure

4.3.2.1.3. Various Subcarrier Allocation Modes

OFDMA supports subchannelization to effectively allocate subchannels for both DL and UL. In WiMax, each user is allocated blocks of subcarriers, not individual subcarriers. This method can decrease the complexity of the subcarrier-allocation algorithm and make the mapping message easier. There are two main types of subcarrier permutations; distributed and adjacent.[50][53] Generally speaking, the distributed subcarrier permutation has very good performance in mobile applications and improves frequency diversity and robustness. Adjacent subcarrier permutation is good at fixed, portable or less mobility surroundings and increases multiuser diversity.

z DL Distributed Subcarrier Permutations: Fully Used Subchannelization (FUSC)

All subchannels and full channel diversity are well utilized in this method by dispensing the allocated subcarriers to subchannels and a permutation mechanism is used here. During

54 transmission, the adjacent cells/sectors may hit each other by certain possibility, because of using the same subcarriers. This mechanism reuses subcarriers to minimize the probability of hits. In addition, fast fading of mobile surroundings would lower the system performance. But frequency diversity can minimize it.

Table 4-5[53] is the summary of subcarrier allocation structure. In DL-FUSC there are two kinds of sets of piloot; variable and fixed sets. The fixed one is used in all OFDM symbols and the variable one is divided into subsets used in odd and even symbols alternatively. It makes the tradeoff of channel estimation appropriate between assigned power and frequency diversity.

Table 4-5 DL Distributed Subcarrier Permutation (FUSC)

z DL and UL Distributed Subcarrier Permutation: Partially Used Subchannelization (PUSC)

In OFDMA specification, the OFDMA DL and UL subframes must begin with DL and UL mode. In DL PUSC, all subchannels will be sliced and assigned to three segments, and it can bee allocated to sectors of the same cell. [53] This method utilizes full-channel diversity by dispensing the allocated subcarriers to subchannels. [53] A permutation mechanism is also used here to minimize the probability of hits and using frequency diversity to minimize the performance degradation.

z Optional DL and UL Adjacent Subcarrier Permutation: Advanced Modulation and Coding (AMC)

Adjacent subcarriers are used by this method to form subchannels. AMC can quickly elect a modulation and coding combination per subchannel, when it is used with fast feedback channel. The subchannels of AMC can use the “water-pouring” types of algorithms, and it can be combined with an AAS option.

Table 4-6[53] is the summary of the AMC subcarrier allocation parameters.

Table 4-6 UL-DL Adjacent Subcarrier Permutation (optional AMC)

4.3.2.2. MAC Layer 4.3.2.2.1. QooS

The functions of the QoS of the IEEE 802.16-2005 include the old four schemes and a new one scheme, extended real-time polling service (ertPS). This scheme is only defined in IEEE 802.16e-2005 MAC-layer. It combines the advantages of UGS and rtPS. This method allocates

56 the uplink bandwidth periodically for MS and with this feature ertPS can coordinate data services which the bandwidth requirements change with time. [50][121][127][128]

4.3.2.2.2. Power-saving

How to extend the battery life is a very important issue of mobility. Therefore power management is a fundamental requirement of a mobile wireless network. The IEEE 802.16e- 2005 defines two modes to operate power saving. [50][129] – [132]

4.3.2.2.2.1. Sleep Mode

This mode is an optional mode of WiMax power management. The MS is defined by two statuses, sleep window and listen window. The sleep window is defined as when a MS temporarily disconnects to the BS for a predetermined period of time. The listen window is following the sleep window, during which the MS reconnects to the BS. The lengths of each sleep and listen window depend on the power-saving classes and is negotiated between the MS and BS. When the MS is in sleep window, this period is called unavailability interval. During the unavailability interval, the MS cannot receive any DL transmission and send any UL transmission. During the availability interval when the MS is in listen window, the MS can receive DL transmission and send UL transmission as normal. BS will not send any DL transmission to MS when it is in sleep window, therefore the MS can turn off one or more physical components for saving power. The BS buffers the SDUs (Service Data Units) for MS and transmits them later to the MS during availability interval.

For different services, there are three power-saving classes:

1. Power-saving class type 1

Class type 1 is used for BE or nrtPS connections. In this class, the listen window size is fixed and followed by a sleep window. Each sleep window size is twice the size of the previous sleep window, but not bigger than the final sleep window size. The size of the initial sleep window and the final sleep window are notified to the MS by the BS before the power-saving class type 1 begins. During the sleep mode, the BS can reset the window size to the initial sleep window size at any time. For UL allocations, the reset process happens under the request from the MS. And for DL allocations, it happens when the BS finds out that the number of listen window is not enough for the traffic.

2. Power-saving class type 2

This class is used for UGS or rtPS connections. The sleep windows and listen windows are all fixed-length and cross each other one by one in order. The BS tells the MS the window size before entering this power-saving mode.

3. Power-saving class type 3

This mode is used for multicast traffic or MAC management traffic. Different from other classes, it has only a single sleep window. Before entering this mode the BS decides the window size and the start time; then notifies the MS. The power-saving operation is inactive, after the sleep window has passed.

4.3.2.2.2.2. Idle Mode

Idle mode is also an optional scheme of WiMax power management, and it can save more power than sleep mode. With this scheme, several BS’s are assigned to a paging group and a MS can move within this area and receive broadcast DL transmission from a BS without switching or registering to the network. When a MS is in idle mode, it listens to the network periodically for determining in which paging group it is and running the paging group update. MS also notifies the network of its appearance while running the paging group update. When a BS needs to connect with the MS in idle mode, all the BSs in the paging group, where the MS is, have to join the paging process. Once the MS receives the paging message, it will end the idle mode and start to connect to the network.

In idle mode, a MS has two statuses;, MS paging-unavailable interval and MS paging-listen interval. When a MS is in MS paging-unavailable interval, it cannot be paged but can power down or run other operations not related with connection with BS, such as scanning. In the MS paging-listen interval, the MS listens to the serving BS for broadcast paging message to see whether it is paged or not. If it is paged, it ends the idle mode and if not, it goes into next MS paging-unavailable interval.

4.4. The future of WiMax

The new member of 802.11, 802.11n, is on the way. 3G is also becoming more popular. Therefore, WiMax definitely has to cooperate with them on a some certain level. WiMax is a wide range wireless communication; Wi-Fi is short range and 3G is also wide range but with narrower bandwidth. Figure 4-7[54] Since they are all wireless communication type systems, the market will be overlapped partially. The services demanded most from end users are e-mail, sharing files, video and/or audio transmission and etc…. [54]

Figure 4-7 Coverage and Capacity for Different Wireless Access Techniques

Basically, the applications of them are different, although they have some features in common. WiMax is developed for MANs; Wi-Fi is for LANs and 3G is the cellular communication standard. Thus, the relationships among them should be of cooperation, not competition. The services of 3G are phone service and limited internet service in a low mobility environment. Wi-Fi provides broadband wireless service in a fixed environment, most likely indoors. WiMax is going to provide a large range broadband wireless service in a fixed or mobility environment. They can cover each other’s drawbacks. [54][133]

In an urban area, IEEE 802.16 has several competitors such as DSL, cable, or fiber. These wired connections have higher speed and are more stable. Most of time, they have already existed when a building was built. Even through wireless communication has been improved a lot; wired communication is still more reliable and safer. The maximum speed that WiMax can reach is about 70 Mbps, but with fiber, it can achieve at least 1 Gbps! Therefore, users would choose wired rather than wireless. [58]

Another problem is the topography. In a highly developed city, these tall buildings would interfere with the transmission of radio. For high frequency microwaves, it requires a clear path to reach the terminal; otherwise the base station must set on a building that is higher than others. On the other hand, lower-frequency microwave using non-line-of-sight (NLOS) technology still has a problem; a signal is difficult to lock on because of many reflections off masonry walls. In

59 an urban area, there are still several problems remaining to be solved for wireless communication. That is the task of WiMax. [58]

In smaller cities and suburban areas, wireless communication has better performance. This is because the competition between wireless and wired is less and the topography is more favorable. It is not worth installing fiber or DSL in this kind of area. Population distribution of these areas is usually scattered, so it will cost more to wire the whole network in this situation. And the numbers of users may be a problem too, thus wireless could be a better choice. Especially when lower-frequency microwave with (NLOS) is more suitable. [58]

Wireless communication has the best performance in rural areas. Not only because of the lack of competition, but also the more friendly geography. It however still has challenges. How to and where to set up stations are two major problems. Distance and population distribution must be considered as well. [58]

When moving in a vehicle, mobile-WiMax can support wireless access. Unlike 3G, mobile- WiMax can provide not only mobility but also BWA. For example, when taking a train for long distance trip, people can use mobile-WiMax to access the Internet for news, video entertainment, or online-game to entertain themselves. However, since the mobile-WiMax is still not commercialized, besides in Korea, the practical operation will continue to be on hold..

CHAPTER FIVE

5. Case study

This chapter will discuss and compare three individual cases using Wi-Fi or WiMax to build up the public wireless network. These cases are located in the cities of Tallahassee, Fl, USA, Taipei, Taiwan and Seoul, South Korea. The network located at Florida State University in Tallahassee, Fl, USA uses 802.11b/g to build an indoor and partial outdoor wireless network. That allows students and faculties to use wireless access in any building. Taipei initiated their public wireless network project in September 2004. This project is contracted by Q-ware communications, Inc. The operation was officially started in December 2005. The service area covers 90% of the population of Taipei City. Installation was officially finished in September 2006 and was named WiFly. South Korea is the first country operating the commercial WiMax network. It is not the only country that uses the fixed-WiMax technology. They are the only country that also combines it with mobile-WiMax. This project is developed by the Korean telecoms industry. The main coverage area is located in its capital, Seoul. The network was launched in June 2006.

Florida State University located in Tallahassee, Florida is the easiest one to gather data, because it is nearby. And by interviewing the related staff [147], more precise and detailed information could be collected. Because the information and business news concerning WiFly are written in English and Chinese, they are easy to read and collect. However because of the time factor, only one interview was conducted concerning WiFly. Most of the data was gathered from the Internet and books. Although it is easy to find the newest information on the Internet, it require cross validation to ensure its validity. This is the most important issue when doing a case study of Wifly. WiBro is the most difficult one of data collection and analysis in case studies. it is because of language barrier. The WiBro official website has an English version, but not for all documents. Most of them are written in Korean. Thus the information is limited and makes the analysis more difficult.

The following sections will introduce these three cases and their implementation. Each section contains analysis and small conclusion. The last section is the overall analysis.

5.1. FSU wireless network 5.1.1. Introduction

The wireless network solution of FSU (Florida State University) is called FSUWIN. WIN means Wireless Integrated Network. This network is based on the existed wired network and provides a more flexible and convenient internet service on campus. With FSUWIN, students

61 and faculties can use their NB’s to access the Internet everywhere on campus. This wireless network follows the IEEE 802.11 standard and covers major buildings and most public areas. Figure 5-1[70]. The department which is in charge of FSUWIN is OTC (Office of Telecommunications and Networking). It is also responsible for maintenance and network security. The first wireless network at FSU was started in 1999, and at that time it was only an experimental network. There was no plan for building a campus-wide network. In 2002, OTC got the funding they needed to install the wireless network at FSU. The goal is to provide an indoor/outdoor wireless public network for all students and faculty. [147]

Figure 5-1 The Coverage of FSU Wireless Network

5.1.2. Implementation

The reason why FSU used WLAN to extend the campus network instead of wired LANs is because it is cheaper and more flexible. The implementation was started from upgrading the existing wired network. The construction and installation were conducted in two parts. First part was Building-to-Building Connectivity (Bridges), and second part was Intra-Building Connectivity (AP). [70] Table 5-1 shows the basic information of FSUWIN.

Part I contains three phases. In phase I, the goal was to install wireless network equipment in 6 buildings and support 89 users. The operators took 4 buildings and created two domains in two connectivity points of the FSU network. After creating these two domains, the network could simply be enlarged by adding additional buildings into each domain. The operators also

62 increased bandwidth, fixed connectivity issues, and hid antennas through upgrading in this phase. The total cost was within $6,000 but if using fiber, it would cost at least $10,000. [70]

In phase II, the operators tracked the outside campus usage and wanted to install wireless network to these small groups or departments. They located the outside campus connections. The goal was to renovate these networks in the FSU extended campus. After evaluating these connections, the operators found out that they could establish connections to these places wirelessly and retrieved many modem lines. Eventually, the operators removed 37 dial-up connections in 5 buildings and then they had more dial-up resources for other users. These buildings used hubs to connect each PC and then used wireless bridges/gateway to connect back to main campus network. In the middle, there is no wired network; it is totally wireless. [70]

In this last phase, operators added 8 additional buildings into the FSU wireless network. The problems of these buildings were that they were either too far away from campus (1.5 miles away) or too expensive to install fiber for each user (spanned city street). By applying the wireless solution, the operators spent around 15,000 dollars without recurring costs or permits. [70]

The Building-to-Building Connections covered 19 buildings with 207 users and 22 bridges. (Document in 09/2006) [70]

Part II contained two phases and mainly focused on the inner connections of a building. In phase I, the operators estimate the areas on the campus where the WLAN was highly needed. These places included three libraries and Student Union. The Student Union is a place where vendors are and students and employees hand-out together. [70]

Phase II was to help the departments which requested an intra-building network to install and setup the FSUWIN network. The process included implementation, integration, optimization and installation. They would make sure the departments did not have any problems when using the WLAN. For achieving maximum efficiency, the operators not only installed and maintained the network, but also evaluated future extensions for the departments. The installation process is documented for future improvement. The departments which have FSUWIN service are the College of Law, Registrar’s Office, College of Business, School of Social-Work, and Mathematics. (Document in 09/2006)[70]

The APs and switches used by FSUWIN are produced by many different companies, such as Vivato Networks, Xirrus and Foundry Networks. The technology details are shown in figure 5- 6[145], figure 5-7[146] And figure 5-8[71]; table 5-2[145], table 5-3[146] and table 5-4[71]. Comparing the Vivato solution to regular AP, it is easy to find that Vivato has longer coverage and could support more users. In the survey “Vivato Switch Evaluation”[70], there are five scenarios had been defined: Fisher, Landis, Shores, Sliger and Stadium. This survey shows that only 1 or 2 pieces of Vivato equipment are needed to cover an area. In the scenario one, Fisher Lecture Hall takes 6 regular APs, three at each side, to cover the whole area, but the Vivato

63 solution needs only one. Figure 5-2[70]. In scenario two, Landis Green needs 3 regular APs, but just one Vivato switch is enough for this area. Figure 5-3[70]. In scenario three, Shores Library total needs five regular APs to cover the whole floors. There are two in the basements, one in the first floor and two in the second floor. In this case Vivato solution needs two switches to cover this area, one in the basement and the other on second floor. Figure 5-4[70]. Scenario four is not available in the survey. The last scenario is Stadium. In this case it takes at least nine regular APs to cover the area and the complex is also high. For Vivato solution, it simply needs only one in the indoor place where there was already power. The Vivato switches are mainly used in the outdoor scenario, and for indoor scenarios, the other types of APs, Xirrus and Foundry Networks, are used.

For ensure the quality of signal strength and maximum capacity, each AP is directly connected to the backbone network; and no more than 20 users for a single AP. If there is a high usage in a certain area, the operators would install more APs to share the throughput. The critical locations of FSUWIN installation were libraries, Student Union and other common public places, because FSUWIN is a public campus network. [147]

Figure 5-2 Fisher Lecture Hall

Figure 5-3 Landis Green

Figure 5-4 Shores Library

Figure 5-5 Stadium

Table 5-1 The Basic Information of FSUWIN

Terms Parameters

Standard IEEE 802.11 a/b/g

Official Initial date 2002

Outdoor 70 ~ 75% Coverage Indoor 10%

Maximum Users per AP 20 (Design Issue)

Average User per Month 2500 ~ 3000

AP model Vivato, Xirrus, Foundry Networks

Data Rate 1 ~ 26 Mbps (actually speed)

Network Type Direct to Backbone network

Table 5-2 Xirrus XS-3700 AP Technology Data

802.11a/b/g Radios 4

802.11a Radios 4

Total Number of Radios 8

Number of Integrated Switch Ports 8

Uplink Ethernet Ports 2 Gigabit

Maximum Wi-Fi Bandwidth 432 Mbps

Integrated Wi-Fi Threat Sensor Yes

Maximum Number of Users per Radio 64

Maximum Number of Users per Array 512

Number of Voice Calls per Array 84

Number of Video Streams per Array 21

Figure 5-6 Xirrus XS-3700 AP

Table 5-3 Foundry Networks IP250 Technology Data

AP And Network Features • Fragmentation Threshold Setting

• Multi-Country Support • Idle Timeout (set at 30 minutes)

• Mounting Bracket for Wall, Ceiling, and • Layer 2 Roaming-802.11f IAPP Tabletop • Integrated AP Locking Mechanism Wireless Data Encryption

• LEDs For Status, Link, Radio a, Radio • RC4 64-bit/128-bit/152-bit Wired b/g Equivalent Privacy (WEP)

• Auto-sensing 10/100BASE-T PoE Port • Temporal Key Integrity Protocol (TKIP)

• Power Over Ethernet (PoE) Support- • Advanced Encryption Standard (AES) 802.3af

• 802.11a, 802.11b, 802.11g Standards • Pairwise Master Key Security Association (PMKSA)

• 802.11a Turbo Mode • WPA2 mixed-mode AES/TKIP

• Simultaneous Dual Radio Support 2.4- Wireless Quality Of Service GHz and 5-GHz

• Integrated Antenna Support For • 802.1p tag mapping to priority queue 802.11a/b/g

• External Antenna Support for • Source/Destination MAC-address 802.11a/b/g mapping to priority queue

• Plenum Rated UL 2043 AP Housing • Ethertype mapping to priority queue

• Automatic Channel Assignment • Spectralink voice prioritization

• Adjustable Power Levels • AP Load Balancing

• AP Load Balancing Radio Support

• Maximum Speed Setting • Simultaneous Dual Radio Support 2.4GHz and 5GHz

• Broadcast/Multicast Rate Limiting • 802.11g protected mode

• VLAN Support-64 VLANs • Automatic channel select

Table 5-3 Cont.

• Automatic Tx power selection

• Maximum station data rate

• Multicast data rate

• Antenna type

• Antenna diversity

• Fragmentation length

Figure 5-7 Foundry Networks IP250

Table 5-4 Vivato 2.4 GHz Indoor & Outdoor Wi-Fi Switch Technology Data

Table 5-4 Cont.

Figure 5-8 Vivato 2.4 GHz Indoor & Outdoor Wi-Fi Switches

5.1.3. Analysis

The FSU wireless network is a very convenient system. It allows all students and faculty to have more freedom to access the Internet without a strict location limitation. The connection is robust and the signal is stable. However there are still complaints and technical problems that exist. Mr. Clint Ringgold, Network Administrator of OTC, said that the primary complaint is about the coverage of the indoor area. At the beginning of the project, the FSUWIN installation was focused on the outdoors, especially for public areas, and had very limited budget. Therefore, if it is not a public area, the installation would be only under the request of the departments. Hence sometimes the coverage is only a half of the building or very small area of the building. That is why users may get a signal on one side of the building, but no signal on the other side of the building. [147]

Concerning this technical problem, Mr. Clint Ringgold mentioned two future goals. One is handoff and the other is the central controller. Mr. Phillip M. Callahan, Assistant Director of OTC, said that installing a central controller is their future direction and it should be the university issue. It is very expensive equipment and the OTC is still waiting for funding. Meanwhile the other issue is to shift the current technology to fit the controller. The function of this controller is to connect to the APs around the campus and monitor the status of each AP, so that the OTC can control the network efficiently. A single controller can connect at least 500 APs. Concerning the other goal, Mr. Clint Ringgold indicated that the current FSUWIN does not support roaming, so users cannot move between two hotspots and maintain a connection. This can be done either by switching to another technology or growing the network. Mr. Ringold described it as a design issue.

The network security is guarded by the FSUWIN login system. Figure 5-9[70]. Every user needs to register a FSUID in order to login to the system to establish a connection. To maintain network reliability, OTC has the right to control the traffic load and RF devices are used for avoiding congestion and interference to wireless signals. The terrain of the FSU campus is favorable to signal transmission, because it does not have high buildings and the density of buildings is low. Therefore multipath fading is not serious to the signal.

Mr. Phillip M. Callahan said that OTC faces several difficulties while installing the network concerning policy and technology issues. Addressing policy, he mentioned ownership and funding. The OTC has skill and management duty, but does not always own the equipment. For example, if a department asks the OTC to install FSUWIN for them and they buy the equipment themselves, then the problem is who is responsible for the cost of maintenance, upgrade and replacement in the future. They need FSU to set a policy for this issue. For technical difficulties, Mr. Clint Ringgold indicated that most of the time there is no data-wires in the location where they want to install an AP, so they need to wire it by themselves or change the planned location. Users complain about the speed and service quality, but it depends on the radio interference, such as distance from the AP, microwave and cell-phone. There is another technical problem that

Mr. Phillip M. Callahan mentioned. He said that sometimes people buy their own devices and just plug them in without reporting them to the OTC. They can cause interference and security problems. He said it may be cheaper to buy an AP in local store, but this would cause security vulnerability in the end.

Figure 5-9 FSUWIN Login System

5.1.4. Conclusion

FSU campus is a relatively small area, so it is suitable for Wi-Fi. The network structure is also designed very well. It is a successful wireless network model. The next step of FSUWIN is to install a central controller, upgrade the network for roaming and extend coverage for more indoor environments. So far, the OTC does not have any plans to upgrade the system to IEEE 802.11n or IEEE 802.16, said Mr. Phillip M. Callahan. FSUWIN does not have many serious problems. The technical problems can be solved by new technologies and the policy issues are dependent on FSU. Beside these issues, FSUWIN is a stable and convenient wireless public network. In the future, if a small town, a village or other smaller areas want to upgrade its existed wired network to wireless public network, FSU is a good example to follow.

5.2. WiFly in Taipei 5.2.1. Introduction

WiFly is part of the Mobile Taiwan Applications Promotion Project. This project started in 2004 and its goal is “to bring out communication industries growth and stronger of national competitiveness based on complete broadband network infrastructure.” [74] WiFly is based on Wi-Fi standard (IEEE 802.11 a/b/g). The WiFly has officially operated for one and a half years

74 since the January 2006. The population coverage is more than 90%, but the geographic coverage is only about 50%. The total number of APs are about 4200.

5.2.2. Implementation

The WiFly infrastructure installation was split in three stages. The first stage was to cover the 30 MRT (Mass Rapid Transit) stations and the surrounding areas, including the underground shopping streets. This stage was finished in May 2005. The second stage included 42 main city arteries, shopping districts, coffee shops, and the rest of MRT stations and so on. There were about 2000 APs that were installed in this stage. In the third stage, there were 9 city hospitals, 53 libraries, 12 Taipei administrative buildings and 600 7-11s (chained convenient stores) included. The final stage was finished in July 2006. More than 4200 APs were used, and up to 90% of the population was covered. This was a brand new world record. [72]

The WiFly uses Nortel Wireless Mesh Network as its fundamental model. Two components, Wireless Gateway 7250 and Wireless Access Point 7220, were used for this model. In this network, the connection between APs is 802.11a and the connection between AP and end users is 802.11b/g. Since there is no need to wire every AP in the network, the cost of installing wired network from gateway to AP and AP to AP can be eliminated. [72]

The Nortel Wireless Mesh Network is a solution of installing the outdoor and indoor Wi-Fi network. It is scalable and secure. An ISP can use it to provide wireless internet service and companies can use it to build inner private networks. The network contains three components; a wireless access point, wireless a gateway and a wireless mesh management platform. Figure 5- 10[75], Figure 5-11[75], Table 5-5[75]

There are two models of APs, Wireless AP 7220 and Wireless AP 7215. The WAP 7220 is an outdoor and indoor-type AP and the WAP 7215 is an indoor-type AP. Figure 5-12[75]. The WAP 7220s can be connected together to cooperate for improving system performance. WAP 7215 only works in indoor locations. It can connect to either wired or wireless networks. WAP 7220 and WAP 7215 also can cooperate together to form the wireless mesh network. The wireless gateway has two models as well; Wireless Gateway 7250 and Wireless Gateway 7240. The WG 7250 can support 10/100 Ethernet interface and 120 WAPs. The WG 7240 also has 10/100 Ethernet interface but only supports 10 WAPs. The Wireless Mesh Management System provides a communication interface for Nortel’s wired and wireless network products. [75]

Figure 5-10 Wireless Mesh Network Structure I

Figure 5-11 Wireless Mesh Network Structure II

Table 5-5 Parameters of WLAN and Mesh Network

Figure 5-12 Wireless Access Point 7220

The advantage of the mesh network is that all APs can share the connections among the network. Therefore the reliability and stability of the network can be improved. If any AP in the network is malfunctioning, the transmission route is not going to fail. Other APs will form an alternative path to keep the network functional. The other advantage is that for places that wired network is difficult to install or cannot be installed; this type of network is a nice alternative solution. [75]

In the AP operation system, Nortel developed a critical scheme called Nortel’s Adaptive Mesh Management Protocol. This protocol can simplify the deployments and optimize the overall performance of the mesh network. There are several important functions in this protocol: auto-discover the neighbor APs, auto-configuration and system synchronization, radio resource management, smart antenna management, dynamic mesh routing, and fast fault recovery. With these technologies, the Nortel Wireless Mesh Network can solve many technical problems such as interference, security, system scalability and reliability, QoS, simplifying installation and lower the cost. [75] Table 5-6[75] is the features of Nortel Wireless Mesh Network.

While dealing with interference, Nortel uses three schemes, Nortel’s Adaptive Mesh Management Protocol, smart antenna mechanism, and dual radio design. The dual radio design uses different spectrums for transmission and reception to avoid self-interference. Nortel’s Adaptive Mesh Management Protocol can sense the status of nearby channels to achieve the maximum usage of channels. With the smart antenna mechanism, data is transmitted in a beamform and directional way, so the data streams do not affect or be affected by other beams. [75]

In the security field, because the wireless network is easier to be attacked by hackers than wired network, to enhance this defect, Nortel’s mesh network adopts WEP/WAP/WAP2 and IPSec (IP Security Protocol). IPSec is derived from VPN (Virtual Private Network) technology. It uses a firewall to filter the incoming data stream before entering the wired network. For extending the network capacity, the mesh network allows the network designers to add new WAPs into the network at any locations in the covered area. This is the scalability of the Nortel Wireless Mesh Network. With the Nortel’s Adaptive Mesh Management Protocol, a new WAP can auto-discover the surrounding WAPs and establish the new optimized transmission path. Similarly, when an AP or more APs are malfunctioning, the Nortel’s Adaptive Mesh Management Protocol can re-route a substitute path to keep the network functional. [75]

Nortel’s Wireless Mesh Network solution can be integrated with WiMax BS. With the indoor/outdoor APs and the mesh network features, this solution gives network designers a lot of freedom to build up a low-cost and network that is also fairly simple. [75]

Table 5-6 Features of Nortel Wireless Mesh Network

Figure 5-13 Nortel Wireless Mesh Network Solution Example [75]

The following section is the data of AP and gateway. [73]

z Wireless Access Point 7220 Technical Specifications Wireless AP 7220 Access Link 802.11b/g (2.4 GHz) Radio System

Center frequency • 2417 MHz to 2457 MHz (i.e., North America) Data rate: 54 Mbps max • Supports 1, 2, 5.5, 11 Mbps (IEEE 802.11b) • Supports 6, 9, 12, 18, 24, 36, 48 and 54 Mbps (IEEE 802.11g) • IEEE 802.11b/g standard rates Access antenna options • Co-linear whip, 5 dBi nominal antenna, SMA connectors • PIFA integrated antenna, 0 dBi nominal SMA connectors Radiated EIRP • +26 dBm typical Receive Sensitivity 802.11b (11Mbps) • -95 dBm typical @ 11 Mbps • -96 dBm typical @ 5.5 Mbps • -98 dBm typical @ 2 Mbps

• -101 dBm typical @ 1 Mbps Receive Sensitivity 802.11g (54 Mbps) • -80 dBm typical @ 54 Mbps • -82 dBm typical @ 48 Mbps • -86 dBm typical @ 36 Mbps • -90 dBm typical @ 24 Mbps • -92 dBm typical @ 18 Mbps • -95 dBm typical @ 12 Mbps • -95 dBm typical @ 9 Mbps • -96 dBm typical @ 6 Mbps

Wireless AP 7220 Transit Link 802.11a (5 GHz) radio system Center frequency • 5740 MHz to 5840 MHz Data rate: 54 Mbps max • Supports 6, 9, 12, 18, 24, 36, 48 and 54 Mbps • IEEE 802.11a standard rates Antenna system gain from radio module card inside the unit • 8.4 dBi nominal Radiated EIRP • +28 dBm typical @ 54 Mbps • +30 dBm typical @ 48 Mbps • +32 dBm typical @ 6-36 Mbps Receive Sensitivity (<1% Packet Error Rate) • -82 dBm typical @ 54Mbps • -85 dBm typical @ 48Mbps • -90 dBm typical @ 36Mbps • -93 dBm typical @ 24Mbps • -98 dBm typical @ 18Mbps • -100 dBm typical @ 12Mbps • -101 dBm typical @ 9Mbps • -101 dBm typical @ 6Mbps Environmental specifications • Operating temperature range: - 40°C min, 50°C max Regulatory • Weather rating: NEMA 4, IP56/Category 2 testing • Safety: UL, CSA • Emissions/radio: FCC Class B, Part 15, RSS 210 Hardware specifications

• Wired network interface: Auto sensing 10/100BaseT Ethernet, 1.5kV surge protection per IEC60950 • Power input nominal: 100V - 240V AC (45Hz – 65Hz) Power consumption • Operating: Indoor or outdoor > 0°C = 8W typical Outdoor < 0°C = 8W – 14W (- 40°C) • Startup: Indoor or outdoor > 0°C = 8W typical Outdoor < 0°C = 24W (short duration) 8W– 14W (- 40°C) Dimensions (without mounting brackets or antennas) • 265mm (10.5 inches) tall x 200mm (8 inches) diameter • Weight: 2.4 kg (5.3 lbs) • Color: Gray Optional accessories • Mounting brackets (right-angle or straight horizontal attachment) • 5m, CAT5 Ethernet indoor/outdoor rated cable for network access point (NAP) operation • Street light photo-electric control power tap 'luminaire' 120/208/240 V • 13dBi, 18dBi and 23dBi TL external antennae

z Wireless Gateway 7250 Technical Specifications

Standard: • 128 MB memory • Processor - 850 MHz • PCI expansion slot - one (available for additional Ethernet interface or hardware accelerator card) • LAN/WAN interfaces Two 10/100 BaseT Ethernet Management/console (DB9) Optional: • One additional 10/100 BaseT Ethernet interface (provides additional LAN/WAN connectivity, if required by network design) • 128 MB RAM upgrade • Hardware encryption accelerator card Physical specifications: • Length: 21 in. (53.3 cm) • Width: 17.25 in. (43.8 cm) • Height: 3.5 in. (8.9 cm) • Weight: 10.0 lb. (4.5 kg)

Operating environment specifications: • Electrical: 90-264 VAC, 2.0 @ 90 VAC, 47-63 Hz • Temperature: 32°-104° F (0°-40° C) • Relative humidity: 10-90% non-condensing

5.2.3. Analysis

The WiFly has been in operation for about 21 months. As of the end of August 2007, there were one hundred and fifty thousand registered customers with an average of six thousand users per month. The population coverage is 96% (two million six hundred thousand people) and the number of APs are 4600. [72] This is the biggest wireless network construction in the world. However, there were many complaints since the first day of service. There are several reasons. First, the signal is unstable with too many dead spots. Second, coverage is concentrated in metropolitan areas; the geographic coverage is only 50%. Third, the signal path is too unpredictable. This is because of the multipath fading and signal reflection. Forth, the price is too high and lacks service applications. These problems can be classified as two parts, mechanical and marketing. The mechanical part includes signal stabilization, AP coverage and security; the marketing part contains service applications and pricing rate. [76] – [78][85]

Compared with the population coverage, the numbers of user is very low. Table 5-7. This project spent a total of approximately thirty million dollars, but ended up facing a low usage situation. Taiwan has a strong manufacturing background and in the 1980s and 1990s, MIT (Made in Taiwan) became a well known mark around the world. With this ability, the infrastructure installation is not a difficult job at all. WiFly is a first city-wide wireless network project; therefore since there is no precedent to learn from, WiFly has to learn through trial and error. [76] – [78][85]

Table 5-7 The Numbers of Subscribers

Terms Registered Users Active Users Industry Customer Date

2006/6 45,000 20,000 N/A

2007/1 110,000 35,000 70

2007/8 150,000 6,000 (monthly) N/A

In the mechanical problem part, the signal is the first problem. It is what most customers complain about. This problem is caused by many factors, such as the number of users, weather, environment, AP locations, and RF interference. This is a common difficulty of all wireless

83 communications. Taipei is a busy city with many office buildings. It is a “concrete jungle”, especially in the metropolitan area. This kind of area is very unfavorable to signal transmission and it will cause serious multipath fading. Thus, the end users may scan the signal, but cannot establish a connection, or even have no signal at all. The other problem is AP. It includes numbers of user and locations. Most of APs are located outdoor, such as wire poles and street lights. Because APs are exposed outside and Taiwan is hot and humid, the failure rate will be high. Moreover these kinds of AP locations also have big security issues. There are 4600 APs in Taipei, management and maintenance is a big challenge. And since the AP installation is focus on population coverage, most suburb residents do not have WiFly signal.

The Q-ware does not limit the numbers of user on an AP. Therefore it may cause congestion and channel overlapping problems. For channel allocation and reallocation, Linda Yeh, the Account Manager in Sales & Customer Service Dep. of Q-ware Systems & Services Corp. [148], said the company has a control platform to manage each AP’s operation channel. In a high density area, if there is congestion or channel overlapping the network engineers can use this platform to switch the operation channel for each AP to release the situation. However, the platform cannot solve the problem completely. In most high usage areas of WiFly, after switching channels, the problem still exists. She said this is because there are only three channels that can be used in one area. It is Wi-Fi’s inherent limitation. [148]

The maximum data rate for each user was 512 kbps for both Uplink and Downlink. In January 2007, Q-ware upgraded to 1Mbps for Downlink and 512 kbps for Uplink. This speed is not a shearing speed, which means it will not be changed even if the users increased. [148] To avoid heavy traffic, Linda indicated that they have a flow control scheme to monitor the traffic flow for APs. The detail of flow control scheme is trade secret, thus she cannot provide more information for it. Channel overlapping and congestion are two major problems of signal transmission. There are several methods to solve the congestion issue, such as, hop-by-hop congestion control and multihop congestion control. [149][150] The main idea of these methods is to send a congestion feedback packet to slow down the source speed to release the congestion. The congestion control has been studied for years. Here is some related researche. [151] – [153]

For improving signal stability, end users can install bridge or high power wireless network cards. These two methods can solve most signal instability. Q-ware also has to install more APs or use high power APs in high usage areas. Since WiMax licenses have been issued in Taiwan. It can also be a solution for signal stability and extending geographical coverage. About security vulnerability, Q-ware said that the wireless connection must be authorized before established and the user account needs two identifications to apply. However there is no one hundred percent security scheme. Users still have to protect their account information and PC from spy-programs. The responsibility is upon both WISP and the users.

The Nortel Wireless Mesh Network solution is a suitable model for WiFly. However, no matter how perfect the model is, Taipei still has its intrinsic limitations, such as high office

84 building density, high population density, humid and hot weather, and unfavorable wired network infrastructure. Because the Nortel Wireless Mesh Network solution adopts mesh structure (APs and Ethernet communicate through Gateways), it decreases the difficulty while installing APs. Nevertheless, the APs’ power supply is still a big issue. And even through APs are not necessary to connect to Ethernet, the Gateways have to. One Gateway can support 120 APs, so 4600 APs need at least 39 Gateways. Considering the connection status, usage and AP locations, it is impossible to use just 39 Gateways to handle the whole network and it will cause serious congestion problem. Therefore, where to install the Gateways and how to wire the Gateways are other important issues. According to Linda Yeh, so far the company does not have an efficient way to solve channel overlapping and congestion problems completely. To install a booster to increase the signal power can be a good solution for these problems, but the National Communications Commission (NCC) has strict rules for signal power. Therefore, Q-ware does not have plans to install it. These problems are still pending. [148]

The marketing problem is another big issue waiting to be solved. When WiFly first came out, it guarantees to provide ubiquitous connection to every citizen. However it was plagued with serious mechanical problems and that is not the only reason for low numbers of users. The WiFly did not meet the needs of end users.

Before WiFly, Wi-Fi had been available in stores for years, such as McDonald’s and some coffee shops. This company was called Yaw Jenq Technology and the service was called “Easy- Up”. At that time the project was cooperated with Intel. It was the first WISP in Taiwan, but ended up in.bankruptcy in May 2005. Easy-Up brought new Internet access to people. This project officially began in May 2003. At that time the wireless interface is NB with built-in centrino platform. With Easy-Up, access internet in McDonald’s or other chain stores was slowly becoming a trend. Currently, sitting in a Starbucks, having a cup of coffee and browsing the Internet has become a very common activity.

With regards to the service policy, Linda Yeh said that the indoor APs are used for mobile devices users and the outdoor APs are mainly for fixed users. The indoor APs are most installed in coffee shops, chain restaurants and other public areas, such as airports. In these places, the signal is more stable and the connection quality is better than outdoor. The outdoor APs are used to provide service to fixed users such as home and industries. These users can use bridges or antennas to receive WiFly signal. However because of many external factors, such as RF interference, weather and transmission path, the signal quality is not as good as the indoor APs. [148] In WiFly’s official website, there is a search engine for WiFly hotspot, but they do not provide AP location search. If a customer wants to know if are there any APs nearby, they need call Q-ware directly. [148]

Wireless Internet access is definitely a business point. WiFly also has the ability to occupy a high market share. But the wrong market positioning and unfriendly pricing rate makes the usage much lower than the prediction. WiFly can never replace the ADSL at home. Therefore it is an

85 auxiliary of extending the ADSL to end users in short range. Unless there is a unique application that only WiFly can do; otherwise keep seeking new service to increase users is still no help. ADSL and WiFly are like roads and bridges, but if WISP tries to use bridges to replace roads; then there is no future for WiFly. Taipei’s city government and Q-ware now are trying to use new VoIP service (Taipei easy call) and WiFly cube to attract more users. [76] – [78][85] However the fundamental problem is why WiFly. There are other choices for indoor wireless internet service, such as Hinet-WLAN and 3G wireless.

It is like selling a game station without games and the price is also way too high. For temporary users, $3.20 for one day access is too expensive; not to mention carrying the NB and looking for signal. WiFly’s unique feature is its ubiquitous connection, but if the feature cannot be used stably and conveniently; then it is still useless.

5.2.4. Conclusion

WiFly tries to provide make people’s lives more convenient. However, it still has fundamental problems to be conquered. How to provide end users more stable connection and higher speed should be the current mission of WiFly. And the customers have to find out whether they need WiFly or not. Taiwanese network market is already saturated and has many competitors, so if WiFly wants to take the market share it will be not easy. Since Wi-Fi has some limitations and the original design is for home use; thus it is difficult to use as the last mile of the network and the city-wide coverage. Since the mother company of WiFly, FAREASTONE, has gotten the operation license of WiMax; the next step of WiFly should be the WiMax infrastructure installation. By adopting WiMax, WiFly may be able to have much better performance and show its real value.

5.3. Wibro in South Korea 5.3.1. Introduction

WiBro (Wireless Broadband Access Service) is a WBA technology. It is a standard extended from 802.16-2004 and 802.16 2005 and uses many techniques the same with them. Figure 5- 14[79]. It has better coverage and stronger mobility than WLAN and higher data rate than 3G. However, 3G has better coverage than WiBro and WLAN has higher data rate than 3G. Therefore WiBro’s position is just in the middle. It can fix their weaknesses, but cannot replace them. This standard was developed by KT (Korean Telecoms Industry) and standardized by TTA (Telecommunications Technology Association) later on.

The manufacturers who are contracted to produce WiBro devices are Samsung, LG, POSDATA and so on. For extending transmission range and decreasing dead spot, the TDD

86 repeater is developing and it will have two version, optical repeater and RF repeater. Figure 5- 15[79] shows the network structure of WiBro. The related works : [134][135]

Figure 5-14 The Relationship Between WiMax and WiBro

Figure 5-15 The Network Structure of WiBro

5.3.2. History

This project was initiated by manufacturers (Samsung and LG) and telecommunication industries (KT and SKT). It started from 2003 to 2005. They total invested approximately thirty six million dollars in this project. The first stage standardization was completed by TTA in late

2004. It was a very big delay since WiMax was well developing at that time. November 2004 was the critical timing for WiBro. Intel started to get involved in WiBro’s development. It cooperated with LG and KT to help develop WiBro’s equipments and ensure that WiBro can be compatible with WiMax. In April 2005, Samsung was elected as a member of WiMax Forum board, so the integration of WiMax and WiBro was speeded up. Therefore, WiBro became a worldwide standard, little by little. U.S. and Japan both had plans to use WiBro to build their WiMax network. In January 2005, three companies got the operation license, KT, HTI (Hanaro Telecom, Inc.), and SKT. And KT and SKT were selected to run the WiBro business. The system installation was initiated in early 2006. The project has three stages. First stage is to finish installation in 20 major cities at the end of 2006. Second is to finish 18 medium scale cities installation in 2007. And finally finish the 46 remote areas’ installation in 2008. [76] – [78][84][85]

5.3.3. Implementation

The operation band of WiBro is between 2.3 GHz to 2.4 GHz. This band is divided into 9 channels. These channels are not overlapped but adjacent and every 3 channels have one guard band, figure 5-16[79] – [82]. WiBro PHY-layer is based on OFDMA for against multipath fading and interference. WiMax supports both TDD and FDD scheme, but WiBro only supports TDD. This is for having biggest spectrum usage. In modulation field, WiBro supports QPSK, 16QAM and 64QAM. And for coding, WiBro uses CTC and other optional method, such as BTC, RS-CC and FEC. Figure 5-17[79] – [82] shows the goal and features of WiBro.

Figure 5-16 The Operation Band

Figure 5-17 WiBro Features

The data rate of Wibro is asymmetric, for a RAS (Radio Access Station) the maximum downlink speed is 18 Mbps and uplink speed is 6 Mbps; for a PSS (Portable Subscriber Station) the maximum downlink speed is 3 Mbps and uplink speed is 1 Mbps. Under mobile conditions, WiBro can only support under 60 kmph movement. It is lower than 3G which is support at least 120 kmph, but much better than Wi-Fi. The coverage of a RAS is defined in three levels, Pico, Micro and Macro. Pico can cover a radius of100 meters area. Coverage of Micro is 400 meters and Macro is 100 meters. This design would make installation of RAS more flexible. [79] – [82]

WiBro also has QoS. It contains three schemes, rtPS, nrtPS and BS; not including UGS. For saving power, WiBro provides sleep mode for end users to extend the battery life. In order to improve the system performance, WiBro introduces scheduling algorithm for transmission service and AMC for coding and modulation. WiBro also supports ARQ/H-ARQ (Hybrid Automatic Request) as well as WiMax. This scheme can effectively re-send the packet, if it is damaged by interference or channel fading. The handoff break time between two RASs is less than 150 ms, it is a very good performance comparing with other wireless algorithm, such as PHS. [79] – [82]

There are still other optional technologies used in WiBro, for example MIMO, AAS and SDCA(Space Time Coding Adoption). The purpose is to extend the coverage and against interference. Table 5-8[79] – [82] contains the PHY specification of WiBro. Figure 5-18[79] – [82] shows the MAC-layer structure.

- Transformation of external network data into MAC SDUs; - Payload header suppression

- System Access - Bandwidth Allocation - Connection set-up - Connection maintenance -QoS

- Authentication - Secure key exchange -Encryption

Figure 5-18 MAC Layer Model

Samsung is the major manufacturer producing the ACRs and RASs. In June 2007, Samsung won the WiMAX World Europe 2007 Award by U-RAS Premium Base Station in System Design. Figure 5-19[83] is the technology data and the picture for U-RAS Premium Base Station. Figure 5-20[83] is the information for ACRs. Table 5-8 is the comparison of WiBro and WiMax.

Figure 5-19 U-RAS Premium

Figure 5-20 ACR Basic Data

Table 5-8 The Comparison of WiBro and WiMax.

Standards WiMax WiBro Terms (IEEE 802.16-2005)

2 – 11 GHz for fixed Operation Band 2.3 – 2.4 GHz 10 – 66 GHz for mobile

1.75 MHz, 3.5 MHz, 7 MHz, Bandwidth 14 MHz, 1.25 MHz, 5 MHz, 10 MHz 10 MHz, 15 MHz, 8.75 MHz

Multiplexing Burst TDM/TDMA/OFDMA OFDMA

Duplexing TDD and FDD TDD

QPSK, 16-QAM, 64- Modulation QPSK, 16-QAM, 64-QAM QAM

Channel Coding CC, CTC, RS, LDPC CTC

UL: 6.1 Mbps Data Rate 1 – 75 Mbps DL: 18.4 Mbps

UGS, rtPS, nrtPS, BS; rtPS, nrtPS, BS; QoS mode ARQ/H-ARQ ARQ/H-ARQ

Power Saving Sleep Mode, Idle Mode Sleep Mode

Coverage Maximum 10 km Urban 1 km

Mobility 120 km/hr 60 km/hr

Other Techs AAS, AMC, MIMO AAS, AMC, MIMO

5.3.4. Analysis

WiBro has been in use for one year. However the number of users is not as high as predicted. At the end of June 2007, there were twenty thousand registered users and the goal is two hundred thousand by the end of 2007. If WiBro wants to reach this goal, it has to increase by 170,000 users in half year! According to a recent study from IDC Korea, the subscribers will reach 130,000 at the end of 2007, 600,000 in 2008, 1.4 million in 2009 and 3.9 million in 2011. If there are various supplemental applications and devices available in the market, the number will grow to 800,000 in 2008 and 5 million in 2011. This prediction may be too optimistic compared to the current situation. [84][85]

In KT’s recent report, 60% of subscribers are between 30 and 40 and 20% are college students in their 20s. They also find that these customers are satisfied with service and price but upset about the prices and availability of devices. KT has to pay subsidization to laptops, WiBro- phone and USB-modem buyers about one hundred to two hundred and sixty dollars. However, after subsidizing, the NB (model # NT-Q35) still costs $1750 to $2400, not including accessories. Until June 2007, the WiBro-related products include two pc-cards, two cell-phones, two mobile- PCs and one notebook. The choices are very limited. [76] – [78][84][85]

The coverage of WiBro so far is still restricted in Seoul and certain nearby areas. The speed is fast but the signal is not stable. Although the mobility is available in a bus or subway, the operation environment is not friendly to use a mobile-PC or NB, not to mention the connection is not stable. People need a seat to operate them. Since 3G also can provide mobile internet service and is more reliable, WiBro really faces a very difficult situation in South Korea. Mobility is the most important feature of WiBro. If subscribers cannot feel the advantages, it would be difficult to increase numbers. Therefore in technical field, WiBro has to not only improve signal quality and coverage, but also the mobility. The device problem, sooner or later, will be solved. It is just a matter of time. When mechanism becomes mature, the price will goes down and products will be more varied.

In marketing field, KT WiBro proposed six applications: Wonder-Eye (Personal Multimedia service, Push-Pull on Demand), Wonder-Media (Video Streaming), Wonder-Media (MMS), Wonder-Phone (Video-Phone), Wonder-Net (Wireless Internet Service), Wonder-Tour (3-D tour of APEC). These applications contain multimedia and internet service. WiBro is an interface between users and network. And it can be found out that these applications are overlapped with 3G and Wi-Fi. Thus WiBro has to provide better system performance; otherwise it is difficult to take marketing share from other two standards. After all, they are not kill applications. For WiBro’s long term development, KT and SKT should put more effort on technical problems.

5.3.5. Conclusion

KT, SKT and Samsung still believe in the future of WiBro. With the extension of coverage and new available devices, the subscribers are going to increase. But prices and service contents are the real problems waiting to be solved. If WiBro is not the essential requirement and there are alternative choices available and better for internet access, WiBro is difficult to increase the market share. Therefore considering the occasions which WiBro is used, mobility and stable signal are the only way for it to keep going.

5.4. Overall analysis

Table 5-9 is the basic information of FSUWIN, WiFly and WiBro. FSU and Taipei use the same standard; Taipei and Seoul belong to the same area. In technical field, Taipei should use WiMax not Wi-Fi, because it is not designed for metropolitan area. Even if WiFly upgrades to the IEEE 802.11n in the future, the overall performance is not going to improve too much. It is like carrying seven people in car with four seats; it can work, but not very well. So the best way is to replace the system to WiMax or combine with WiMax. In Seoul, WiBro’s major problem is mobility and it is also the selling point. Using WiMax to cover a city is a suitable way. It is what WiMax designed for. Even though WiBro’s overall performance is not as good as expected so far, the technical issues are solvable by new devices and technologies. As for cost and service problems, it is the same road and bridge issues with WiFly. Both WiFly and WiBro are WISPs and the best way for them to survive in future Internet market is to integrate with ISPs.

FSUWIN is the most successful case within these three cases. In the future, after installing the central controller, OTC can easily monitor the network status and provide more reliable service. FSUWIN reached its goal and does a good job. One more factor of FSUWIN’s success is no financial burden. It is not a commercial network and is supported mainly by tuition. Thus it does not have to worry about users and application issues. This feature makes its operation simpler than other two systems. In short, it is a good solution for smaller and local area.

So far WiFly and WiBro are still seeking their marketing positioning. 3G, Wi-Fi, WiMax and wired internet service where the balance point for these technologies is an important issue to industries, manufactures and researchers. Table 5-10 is the basic information for internet access methods. The service providers and WISP should not only focus on providing content but also have to think about the demand of wireless network users.

Based on the patterns of people’s behaviors accessing the Internet, these technologies should be coexisted. Figure 5-21 is an example for using these technologies. When a man wants to go to a state park and has the Internet access all the way, he has to bring his phone and NB. At home he can use Wi-Fi to collect the information of the state park. During transportation, he can access the Internet by mobile-WiMax. And then, when he arrives the state park, he can still check his e- mail by fixed-WiMax service. If he wants to take a rest, he can stay in the state park service center where Wi-Fi is available. Moreover, all the way, he can make a video phone call anywhere at any time. It is one possibility for future development.

WiFly and WiBro were just commercialized. They have to learn from mistakes, because they are the pioneers of public wireless network. The technical issues they have right now are all solvable. WiBro can showed its real capacity only after the mobility is truly functional and devices price is acceptable. WiFly can adopt WiBro’s technologies to improve its signal and coverage issues. The wireless city story has not finished yet, and WiFly and WiBro still have a chance for success.

Figure 5-21 An Example Application of 3G, Wi-Fi and WiMax

5.5. Other developing wireless network

In U.S., Sprint Nextel has announced to cooperate with Samsung, Motorola and Intel will build a national-wide 4G broadband mobile network. They are going to invest $1 billion in 2007 and between $1.5 billion and $2 billion in 2008. In China, they are developing their own 4G standard, McWill. It is not an international standard and has a worse performance than WiMax. The Chinese government set many restrictions on WiMax and wants to develop their own standard, so in short term, WiMax is not going to develop in China. In Europe, England started to install fixed-WiMax since 2005. There are two companies in the projects, Telebria and Community Internet (CI). Japan issued three WiMax licenses at the end of 2006. They are KDDI, NTT DoCoMo and Yozan. Taiwan issued six WiMax licenses and split into two area, North district and South district. The test service will be launched in 2008.

Table 5-9 The Basic Data of FSUWIN, WiFly and WiBro

Service Name FSUWIN WiFly WiBro Parameters

Standard IEEE 802.11 b/g IEEE 802.11 a/b/g IEEE 802.16e

FSU, Tallahassee, Taipei Seoul Location FL, USA Taiwan South Korea

Operator FSU, OTC Q-ware KT, SKT, Samsung

Launch date 2003 ~ 2004 January 2006 July 2006

Band 2.4 GHz ISM 2.4 GHz ISM 2.3 ~ 2.4 GHz

UL:6.1 Mbps Data rate ~ 25 Mbp UL/DL:512 kbps DL:18.4 Mbps

Indoor: 600 m Coverage N/A Urban ~ 1 km Outdoor: 7200 m

Overall Coverage 75% of campus 90% of population N/A

Mobility X X 60 km/hr

Hand-Off Ability X X Low

All students and Subscribers 150,000 (6/2007) 24,000(7/2007) faculty

Users per device 100 N/A N/A

Vivato, Xirrus, AP model Foundry Nortel AP 7220 Samsung RAS Networks

Operation Campus Metropolitan area Metropolitan area environment

N/A Internet access WiFly cube Applications (VoIP will not be FSU campus life Taipei easy call supported)

Table 5-10 Current Internet Access Technologies

Name Wired 3G Wi-Fi WiMax Parameters Internet

WCDMA IEEE 802.11 IEEE 802.16 Standard Ethernet CDMA2000 a/b/g/n d/e

WiMax Supporter ITU Wi-Fi Alliance N/A Forum

2 – 11 GHz 2100MHz ~ for fixed Band 2.4 GHz N/A 2200MHz 10 – 66 GHz for mobile

Maximum Coverage 1 ~ 5 km 100m ~1000m Largest 10 ~ 50 km

Hand-Off Highest Low Low N/A Ability

54 Mbps (a/g) 100 Mbps ~ Data rate 384 kbps 70 Mbps 200Mbps (n) 1Gbps

Mobility 100 km/hr < 20 km/hr 120 km/hr N/A

Cell-phone/ Devices Cell-phone PC/NB PC/NB PC/NB

FOMA (Japan), WiBro Current WiFly (Taipei, JUNE (Korea), (Seoul, ADSL/Cable application Taiwan) i-mode (Taiwan) Korea)

CHAPTER SIX

6. Conclusion

Wireless communications have been available for more than ten years. Today, 3G, Wi-Fi and WiMax are the most popular and widely discussed standards. Telecommunication industries hope that 3G can bring them new fortune, because 3G has improved enough to support e-mail and small data download. Therefore telecommunication industries want to occupy more of the market share of wireless Internet access by using 3G technology. But compared to Wi-Fi and WiMax, it is still too slow. Moreover, with the new coming Wi-Fi standard, IEEE 802.11n, WLAN becomes more powerful than before and can overcome many past limitations, such as coverage, stability and speed. It is slowly catching up with the capacities of wired networks. A Wired network’s advantage is its reliability, but its biggest disadvantage is path design. In a city, to wire a line to a house or a building is easy; however in remote or rural areas it is a huge issue. Thus WiMax is emerging as a possible solution.

Table 6-1 3G, Wi-Fi and WiMax Overall Comparison

Standard 3G Wi-Fi WiMax Parameters

Coverage (a single tower) 2 (< 1 km) 3 1 (1 ~ 5 km)

Data Rate 3 (< 300 kbps) 1 (~ 25Mbps) 2

Throughput 3 2 1

Mobility and hand-off 1 3 (WiFly) 2 (WiBro) 1: Very good, 2: Good, 3: Bad

Table 6-1 is a comprehensive result of Table 5-11 and 5-12. 3G, Wi-Fi and WiMax. They have individual and unique fundamental technologies, but their capacities are partially overlapped. 3G can be a way to access the Internet and Wi-Fi and WiMax can provide VoIP service as an alternative for the traditional phone-call. However in Table 6-1, it can be found that for Internet service, 3G is too slow for real-time service, such as downloading larger files or watching online-TV. For phone-call service, VoIP’s quality is not as good as 3G; it is because of the design issue. 3G is telecommunication technology and Wi-Fi and WiMax are Internet technologies. They were designed for different purposes. Even though these technologies are overlapped in the application field, it is impossible for them to replace each other at this point. The overlap can be treated simply as a backup. It is better to let a cell-phone just be a cell-phone.

WiFly used Wi-Fi to cover 60% of the city including all metropolitan areas. It takes 4200 APs to do it. If it is replaced by WiMax, it will not take that much. And for FSU campus, if WiMax is applied, just one or two base stations are enough to cover the whole campus. However, two towers are not enough to take care of the daily usage. Moreover, it is always difficult to receive a signal from the outside in a concrete building. Therefore it needs a bridge or a gateway to direct a signal into a building. Wi-Fi may have low coverage, but it can be fixed by using more APs in a hotspot. The advantage is that it can not only extend the coverage but also reduce the flow rate for each AP.

To build a city-wide wireless network is going to be a trend and it is feasible, but if a network only uses Wi-Fi or WiMax, it would suffer cost, coverage, signal and mobility issues. The best way would be for them to be used in combination with each other. So far there is no single technology can feed all different needs. In local area Wi-Fi has strong capability and WiMax is designed as the extension of ADSL or cable. The combination of these two standards can form a most efficient wireless network. Therefore the future of the wireless technologies should be created with the cooperation and integration of these technologies.

There is a new standard IEEE 802.11u is based on this idea. It tries to make Wi-Fi interwork with other different types of networks. [154] The IEEE 802.11u allows devices to interwork with external networks. For this purpose, interworking refers to MAC layer enhancements. It allows the higher layer functionality to provide the overall end to end solution. Instead of telling it what to do, the IEEE 802.11u only helps upper layers to establish an end to end connection with external networks. It provides a “virtual point of presence” for many different networks via a single AP. IEEE 802.11u assists the advertising and connection to remote service beyond the DS and provides information to the STA about the external network prior to association. In fact, the IEEE 802.11u is an amendment to the IEEE 802.11 standard. The final approval of IEEE 802.11u is expected to be September 2009. [154] the Figure 6-1 shows the possible implementation of the future network.

Figure 6-1 Integration of WiMax and Wi-Fi

100

BIBLIOGRAPHY [01]. IEEE 802.11 official website (http://www.ieee802.org/11/) [02]. Vaughan-Nichols, S.J, "Will the New Wi-Fi Fly?", Computer, Volume 39 , Issue 10, page(s): 16 - 18, Oct. 2006 [03]. Reinwand, Cole C.; "Municipal Broadband - The Evolution of Next Generation Wireless Networks", Radio and Wireless Symposium, 2007 IEEE, Pages: 273 - 276, Jan. 2007 [04]. IEEE Std 802.11-1997 (http://standards.ieee.org/getieee802/802.11.html) [05]. IEEE Std 802.11-1999 (http://standards.ieee.org/getieee802/802.11.html) [06]. IEEE 802.11b-1999 (http://standards.ieee.org/getieee802/802.11.html) [07]. IEEE 802.11a-1999 (http://standards.ieee.org/getieee802/802.11.html) [08]. Kapp, S.; "802.11a. More bandwidth without the wires", Internet Computing IEEE, Volume 6 , Issue 4, page(s): 75 - 79, July-Aug. 2002 [09]. Doefexi, A.; Armour, S.; Beng-Sin Lee; Nix, A.; Bull, D.; "An evaluation of the performance of IEEE 802.11a and 802.11g wireless local area networks in a corporate office environment", Communications, 2003. ICC '03. IEEE International Conference on, Volume 2, page(s): 1196 - 1200, May 2003 [10]. IEEE 802.11g-2003 (http://standards.ieee.org/getieee802/802.11.html) [11]. Clincy, V.; Sitaram, A.; Odaibo, D.; Sogarwal, G,; "A Real-Time Study of 802.11b and 802.11g", Systems and Networks Communication, 2006. ICSNC '06. International Conference on, page(s): 69 - 69, Oct. 2006 [12]. Intel Corporation (http://www.intel.com/) [13]. Sunghyun Choi; del Prado Pavon, J.; "802.11g CP: a solution for IEEE 802.11g and 802.11b inter-working", Vehicular Technology Conference, 2003. VTC 2003-Spring. The 57th IEEE Semiannual, Volume 1, page(s): 609 - 694, April 2003 [14]. Wang, S.-C.; Chen, Y.-M.; Tsern-Huei Lee; Helmy, A; "Performance evaluations for hybrid IEEE 802.11b and 802.11g wireless networks", Performance, Computing, and Communications Conference, 2005. IPCCC 2005. 24th IEEE International, page(s): 111 - 118, April 2005 [15]. Mahasukhon, P.; Hempel, M.; Ci, S.; Sharif, H.; “Comparison of Throughput Performance for the IEEE 802.11a and 802.11g Networks”, AINA, pp. 792-799, 21st International Conference on Advanced Networking and Applications (AINA '07), 2007 [16]. Wei-Min S. Studio, Book, "Networking Essentials", 2006, In Chinese. [17]. Gast S. M., Book, “802.11 Wireless Networks: The Definitive Guide”, Second Edition, 2005 [18]. Heusse, M.; Rousseau, F.; Berger-Sabbatel, G.; Duda, A.; "Performance anomaly of 802.11b", INFOCOM 2003. Twenty-Second Annual Joint Conference of the IEEE Computer and Communications Societies. IEEE, Volume 2, page(s): 836- 843, April 2003 [19]. Paruchuri, R.C.; Agrawal, P.; "Interference Study of 802.11b Networks for Proactive Performance Management", Network Operations and Management Symposium, 2006. NOMS 2006. 10th IEEE/IFIP, 2006 [20]. Vassis, D.; Kormentzas, G.; Rouskas, A.; Maglogiannis, I.; "The IEEE 802.11g standard for high data rate WLANs", Network, IEEE, Volume 19, Issue 3, June 2005 [21]. 802.11 PHY Layers

101

(http://searchmobilecomputing.techtarget.com/searchMobileComputing/downloads/CWA P_ch8.pdf) [22]. Crow, B.P.; Widjaja, I.; Kim, L.G.; Sakai, P.T.; "IEEE 802.11 Wireless Local Area Networks", Communications Magazine, IEEE, Volume 35, Issue 9, Sept. 1997 [23]. Qi, Y.; Li, H.; Hara, S.; Kohno, R.; "Clear Channel Assessment (CCA) with multiplexed preamble symbols for impulse Ultra-wideband (UWB) communications", Ultra- Wideband, The 2006 IEEE 2006 International Conference on, Sept. 2006 [24]. Vermeer, V.; "Wireless LANs; why IEEE 802.11 DSSS?", WESCON/97. Conference Proceedings, Nov. 1997 [25]. Spasojevic, Z.; Burns, J.; "Performance comparison of frequency flopping and direct sequence spread spectrum systems in the 2.4 GHz range", Personal, Indoor and Mobile Radio Communications, 2000. PIMRC 2000. The 11th IEEE International Symposium on, Volume 1, Sept. 2000 [26]. Wiegandt, D.A.; Zhiqiang Wu; Nassar, C.R.; "High-performance, high-throughput IEEE 802.11 DSSS WLAN via carrier-interferometry chip-shaping", Broadband Communications for the Internet Era Symposium digest, 2001 IEEE Emerging Technologies Symposium on, Sept. 2001 [27]. Barman, K.; Malipatil, A.V.; "ICI equalizer in a CCK based DSSS communication system", TENCON 2003. Conference on Convergent Technologies for Asia-Pacific Region, Volume 4, Oct. 2003 [28]. Chang-Joo Kim; Hyuck-Jae Lee; Hwang-Soo Lee; "Adaptive acquisition of PN sequences for DSSS communications", Communications, IEEE Transactions on, Volume 46 , Issue 8, Aug. 1998 [29]. Roshan, Pejman., Book, “802.11 Wireless LAN fundamentals”, 2004 [30]. Singh, G.; Alphones, A.; "OFDM modulation study for a radio-over-fiber system for wireless LAN (IEEE 802.11a)", Information, Communications and Signal Processing, 2003 and the Fourth Pacific Rim Conference on Multimedia. Proceedings of the 2003 Joint Conference of the Fourth International Conference on, Volume 3, Dec. 2003 [31]. Shenghao Yang; Yuping Zhao; "Channel estimation method for 802.11a WLAN with multiple-antenna", Communications, 2004 and the 5th International Symposium on Multi-Dimensional Mobile Communications Proceedings. The 2004 Joint Conference of the 10th Asia-Pacific Conference on, Volume 1, 29 Aug.-1 Sept. 2004 [32]. Wiegandt, D.A.; Nassar, C.R.; "High-performance 802.11a wireless LAN via carrier- interferometry orthogonal frequency division multiplexing at 5 GHz", Global Telecommunications Conference, 2001. GLOBECOM '01. IEEE, Volume 6, 25-29 Nov. 2001 [33]. Hongwei Yang; "A road to future broadband wireless access: MIMO-OFDM-Based air interface", Communications Magazine, IEEE, Volume 43, Issue 1, Jan. 2005 [34]. Xianbin Wang; Wu, Y.; Chouinard, J.-Y.; Hsiao-Chun Wu; "On the design and performance analysis of multisymbol encapsulated OFDM systems", Vehicular Technology, IEEE Transactions on, Volume 55, Issue 3, May 2006 [35]. Gerakoulis, D.; Salmi, P.; "An interference suppressing OFDM system for ultra wide bandwidth radio channels", Ultra Wideband Systems and Technologies, 2002. Digest of Papers. 2002 IEEE Conference on, 21-23 May 2002 [36]. Zou, W.Y.; Yiyan Wu; "COFDM: an overview", Broadcasting, IEEE Transactions on, Volume 41, Issue 1, March 1995

102

[37]. Sun, Y.; "Bandwidth-efficient wireless OFDM", Selected Areas in Communications, IEEE Journal on, Volume 19, Issue 11, Nov. 2001 [38]. Maharatna, K.; Grass, E.; Jagdhold, U.; "A 64-point Fourier transform chip for high- speed wireless LAN application using OFDM", Solid-State Circuits, IEEE Journal of, Volume 39, Issue 3, March 2004 [39]. Ye Li; Chuang, J.C.; Sollenberger, N.R.; "Transmitter diversity for OFDM systems and its impact on high-rate data wireless networks", Selected Areas in Communications, IEEE Journal on, Volume 17, Issue 7, July 1999 [40]. Zhengdao Wang; Giannakis, G.B.; "Complex-field coding for OFDM over fading wireless channels", Information Theory, IEEE Transactions on, Volume 49, Issue 3, March 2003 [41]. Agrawal, D.; Tarokh, V.; Naguib, A.; Seshadri, N.; "Space-time coded OFDM for high data-rate wireless communication over wideband channels", Vehicular Technology Conference, 1998. VTC 98. 48th IEEE, Volume 3, 18-21 May 1998 [42]. Zhiqiang Liu; Yan Xin; Giannakis, G.B.; "Space-time-frequency coded OFDM over frequency-selective fading channels", Signal Processing, IEEE Transactions on [see also Acoustics, Speech, and Signal Processing, IEEE Transactions on], Volume 50, Issue 10, Oct. 2002 [43]. IEEE 802.11h-2003 (http://standards.ieee.org/getieee802/802.11.html) [44]. IEEE 802.11j-2004 (http://standards.ieee.org/getieee802/802.11.html) [45]. IEEE 802.16 (http://www.ieee802.org/16/) [46]. IEEE 802.16a (http://www.ieee802.org/16/) [47]. IEEE 802.16c (http://www.ieee802.org/16/) [48]. IEEE 802.16-2004 (http://www.ieee802.org/16/) [49]. IEEE 802.16e-2005 (http://www.ieee802.org/16/) [50]. Andrews. J. G., Ghosh A., Muhamed R., Book, “Fundamentals of WiMax”, 2007 [51]. Liu H., Li G., Book “OFDM-Based Broadband Wireless Networks ”, 2005 [52]. Johnston D., Walker, J., “Overview of IEEE 802.16 security”, Security & Privacy Magazine, IEEE, Vol. 02, Issue 3, May-June 2004 [53]. Yaghoobi H., "Scalable OFDMA Physical Layer in IEEE 802.16 WirelessMAN.", Intel Technology Journal. Vol. 8, issue 3, 2004 [54]. Liangshan Ma, Dongyan Jia, “The Competition and Cooperation of WiMAX, WLAN and 3G”, Mobile Technology, Applications and Systems, 2005 2nd International Conference on 15-17 Nov. 2005 Page(s):1 – 5 [55]. Tanenbaum A. S., Book “Computer Network”, 4th edition, 2002 [56]. Borisov N., Goldberg I., Wagner D., "Intercepting Mobile Communications: The Insecurity of 802.11" , http://www.isaac.cs.berkeley.edu/isaac/wep-faq.html [57]. Borisov N., Goldberg I., Wagner D., "Intercepting mobile communications: the insecurity of 802.11." MOBICOM 2001, pp180–189. [58]. Sweeney D., Book “WiMax operator's manual : building 802.16 wireless networks ”, 2006 [59]. WiMAX Forum - WiMAX Home < http://www.wimaxforum.org/home/> [60]. IEEE 802.16 http://www.ieee802.org/16/ [61]. Xiao Y., “IEEE 802.11n: enhancements for higher throughput in wireless LANs”, Wireless Communications, IEEE [see also IEEE Personal Communications], Volume 12, Issue 6 ,Dec. 2005

103

[62]. Lorincz, J. Begusic, D., “Physical layer analysis of emerging IEEE 802.11n WLAN standard”, Advanced Communication Technology, 2006. ICACT 2006. The 8th International Conference, Volume 1, 20-22 Feb. 2006 [63]. Abraham, S. Meylan, A. Nanda, S., “802.11n MAC design and system performance”, Communications, 2005. ICC 2005. 2005 IEEE International Conference on, Volume 5, 16-20 May 2005 [64]. Cai, L.X. Xinhua Ling Xuemin Shen Mark, J.W. Hua Long, “Capacity Analysis of Enhanced MAC in IEEE 802.11n”, Communications and Networking in China, 2006. ChinaCom '06. First International Conference on, 25-27 Oct. 2006 [65]. 簡永懿(Yung Yih Jian) , “IEEE 802.11n 技術提案摘之研究－PHY Layer Frame Format”, May 3 2005, Information and Communications Research Labs, Industrial Technology Research Institute, in Chinese. [66]. 郭俊均(Chun Chun Guo), “IEEE 802.11n 合併提案之研究－目前合併提案中 PHY 的內容與分析 (IEEE 802.11n merge study－PHY study and analysis)”, December 15 2005, Information and Communications Research Labs, Industrial Technology Research Institute, in Chinese. [67]. IEEE 802.3 (http://www.ieee802.org/3/) [68]. ITU (http://www.itu.int/net/home/index.aspx [69]. Jhih-Jiang J., Book, “3rd Generation Mobile Telecommunication Technology”, 2006, In Chinese. [70]. FSU - OTC (http://otc.fsu.edu/Networking/NetMenu.html) [71]. Vivato Networks Inc. (http://www.vivato.net/) [72]. WiFly (http://www.wifly.com.tw/Wifly6/tw) [73]. Nortel Networks (http://www.nortel.com/) [74]. M-Taiwan (http://www.mtaiwan.org.tw/mp.asp?mp=3) [75]. Nortel Networks, “Solution Brief Wireless Mesh Network”, (http://media.govtech.net/Digital_Communities/Nortel/WirelessMesh_SolutionBrief.pdf) [76]. CNET, Business News, In Chinese, (http://taiwan.cnet.com/) [77]. DIGITALWALL, Business News, In Chinese, (http://www.digitalwall.com/) [78]. DIGITIMES, Business News, In Chinese, (http://www.digitimes.com.tw/) [79]. WiBro (http://www.wibro.or.kr/) [80]. Soon-Young Y., “Introduction to Wibro Technology”, Telecom R&D Center Samsung Electronics Co, Ltd., September 2004, (http://www.itu.int/ITU-D/imt-2000/documents/Busan/Session3_Yoon.pdf) [81]. Balaji S. H., Manoj C., Taewon K, “South Korea Plans for ‘WiBro’”, Samsung Electronics Co., December 2004 (http://www.cewit.org.in/docms/ws3/Wimaxwsbalajih.pdf) [82]. MIC, “Wireless Internet Access : WiBro”, Korea, July 2005 (http://www.telecomnetworks.ru/datadocs/doc_350tu.pdf) [83]. Samsung Electronics Co. (http://www.samsung.com/us/) [84]. WiMax.com, Business News (http://www.wimax.com/) [85]. AsiaMedia, Business News (http://www.asiamedia.ucla.edu/) [86]. Wikipedia (http://www.wikipedia.org/) [87]. Sung Ki P.; Pyeong-Jung S.; Geun Sik B.; “Joint optimization of radio repeater location and linking in WLL systems with 2.3 GHz frequency band”, Communications, 1999. ICC '99. 1999 IEEE International Conference on, Volume 3, page(s): 1617 – 1621, June 1999

104

[88]. Mineno T.; Babji T.,"Issues and challenges of implementing a wireless local loop (WLL) based telephone access network", Personal Wireless Communications, 1997 IEEE International Conference on, page(s): 366 - 370, Dec. 1997 [89]. Noerpel A.R.; Yi-Bing L., "Wireless local loop: architecture, technologies and services", Personal Communications, IEEE [see also IEEE Wireless Communications], Volume 5, Issue 3, page(s): 74 - 80, June 1998 [90]. Haiying Z.;Boucher L.M.; Wachira M., "Technology comparisons in wireless local loop", Personal, Indoor and Mobile Radio Communications, 2001 12th IEEE International Symposium on, Volume 1, page(s): A-46 - A-50, Oct 2001 [91]. Seidel S.Y., "Radio propagation and planning at 28 GHz for local multipoint distribution service (LMDS)", Antennas and Propagation Society International Symposium, 1998. IEEE, Volume 2, page(s): 622 - 625, June 1998 [92]. Schellenberg J., "Designing the digital MMDS network for maximum benefit: technological considerations in service, coverage and subscriber issues", Broadcasting Convention, 1997. IBS 97., International (Conf. Publ. 447), page(s): LP1 - LP5, Sept. 1997 [93]. Ghosh A.; Wolter D.R.; Andrews J.G.; Chen R., "Broadband wireless access with WiMax/802.16: current performance benchmarks and future potential", Communications Magazine, IEEE, Volume 43, Issue 2 page(s): 129- 136, Feb. 2005 [94]. Eklund C.; Marks R.B.; Stanwood K.L.; Wang, S., "IEEE standard 802.16: a technical overview of theWirelessMANTM air interface for broadband wirelessaccess", Communications Magazine, IEEE, Volume 40, Issue 6, page(s): 98-107, Jun 2002 [95]. Dong-Hoon C.; Jung-Hoon S. ;Min-Su K.; Ki-Jun H., "Performance analysis of the IEEE 802.16 wireless metropolitan area network", Distributed Frameworks for Multimedia Applications, 2005. DFMA '05. First International Conference on, page(s): 130- 136, Feb. 2005 [96]. Carl E., "The IEEE 802.16 Standard for Broadband Wireless Access" (http://www2.ing.unipi.it/ew2002/proceedings/pmp002.pdf) [97]. Schwingenschlogl C.; Dastis V.; Mogre P.S.; Hollick M.; Steinmetz R., "Performance Analysis of the Real-time Capabilities of Coordinated Centralized Scheduling in 802.16 Mesh Mode", Vehicular Technology Conference, 2006. VTC 2006-Spring. IEEE 63rd, Volume 3, page(s): 1241 - 1245, 2006 [98]. Nuaymi L.; Noun Z., "Simple Capacity Estimations in WIMAX/802.16 System", Personal, Indoor and Mobile Radio Communications, 2006 IEEE 17th International Symposium on, page(s): 1 - 5, Sept. 2006 [99]. Juan L.; Haggman S. G., "Performance of IEEE802.16-2004 Based System in Jamming Environment and its Improvement with Link Adaptation", Personal, Indoor and Mobile Radio Communications, 2006 IEEE 17th International Symposium on, page(s): 1 - 5, Sept. 2006 [100]. Christian H., "Analysis and performance evaluation of the OFDM-based metropolitan area network IEEE 802.16", European Wireless 2004 Conference, Volume 49, Issue 3, Pages 341-363, 19 October 2005 [101]. Bo H.; Weijia J.; Lidong L., "Performance evaluation of scheduling in IEEE 802.16 based wireless mesh networks", Computer Communications, Volume 30, Issue 4, Pages 782-792, 26 February 2007

105

[102]. Koffman I.; Roman V., "Broadband wireless access solutions based on OFDM access in IEEE802.16", Communications Magazine, IEEE, Volume 40, Issue 4, page(s): 96-103, Apr 2002 [103]. Baxley R.J.; Kleider J.E.; Zhou G.T., "Pilot Design for IEEE 802.16 OFDM and OFDMA", Acoustics, Speech and Signal Processing, 2007. ICASSP 2007. IEEE International Conference on, Volume 2, page(s): II-321 - II-324, April 2007 [104]. Yanxin Y.; Tomisawa M.; Yi G.; Yong Liang G.; Gang W.; Choi Look L., "Joint timing and frequency synchronization for IEEE 802.16 OFDM systems", Mobile WiMAX Symposium, 2007. IEEE, page(s): 17 - 21, March 2007 [105]. Al-Gharabally M.; Das P., "Performance analysis of OFDM in frequency selective time- variant channels with application to IEEE 802.16 broadband wireless access", Volume 1, page(s): 351 - 356, Oct. 2005 [106]. Jisang Y.; Kanghee K.; Kiseon K., "Capacity evaluation of the OFDMA-CDMA ranging subsystem in IEEE 802.16-2004", Wireless And Mobile Computing, Networking And Communications, 2005. (WiMob'2005), IEEE International Conference on, Volume 1, Page(s):217 - 223, 2005 [107]. de Moraes L.F.M.; Maciel P.D.Jr.; "Analysis and evaluation of a new MAC protocol for broadband wireless access", Wireless Networks, Communications and Mobile Computing, 2005 International Conference on, Volume 1, page(s): 107- 112, June 2005 [108]. GuoSong C.; Deng W.; Shunliang M., "A QoS architecture for the MAC protocol of IEEE 802.16 BWA system", Communications, Circuits and Systems and West Sino Expositions, IEEE 2002 International Conference on, Volume 1, Page(s):435 - 439, 29 June-1 July 2002 [109]. Haitang W.; Bing H.; Agrawal, D.P., "Admission Control and Bandwidth Allocation above Packet Level for IEEE 802.16 Wireless MAN", Parallel and Distributed Systems, 2006. ICPADS 2006. 12th International Conference on, Volume 1, 12-15 July 2006 [110]. Cicconetti C.; Erta A.; Lenzini L.; Mingozzi E., "Performance Evaluation of the IEEE 802.16 MAC for QoS Support", Mobile Computing, IEEE Transactions on, Volume 6, Issue 1, Page(s):26 - 38, Jan. 2007 [111]. Lin J.; Sirisena H.; "Quality of Service Scheduling in IEEE 802.16 Broadband Wireless Networks", Industrial and Information Systems, First International Conference on, Page(s):396 - 401, 8-11 Aug. 2006 [112]. Sun J.; Yanling Y.; Hongfei Z.; "Quality of Service Scheduling for 802.16 Broadband Wireless Access Systems", Vehicular Technology Conference, 2006. VTC 2006-Spring. IEEE 63rd, Volume 3, Page(s):1221 - 1225, 2006 [113]. Soong A.C.K., "Technological enablers of the IEEE wireless metropolitan area network (IEEE 802.16)", Quality of Service in Heterogeneous Wired/Wireless Networks, 2005. Second International Conference on, 22-24 Aug. 2005 [114]. Piechoclri, R.J.; Tsoulos, G.V., "A macrocellular radio channel model for smart antenna tracking algorithms", Vehicular Technology Conference, 1999 IEEE 49th, Volume 3, Page(s):1754 - 1758, 16-20 May 1999 [115]. Hamid Z.; Khan Shoab A., "An Augmented Security Protocol for WirelessMAN Mesh Networks", Communications and Information Technologies, 2006. ISCIT '06. International Symposium on, Page(s):861 - 865, Oct. 18 2006-Sept. 20 2006 [116]. Sen X.; Manton M.; Chin-Tser H., "Security issues in privacy and key management protocols of IEEE 802.16", ACM Southeast Regional Conference, Pages: 113 - 118, 2006

106

[117]. Yun Z.; Yuguang F., "Security of IEEE 802.16 in Mesh Mode", Military Communications Conference, 2006. MILCOM 2006, Page(s):1 - 6, 23-25 Oct. 2006 [118]. Sen X.; Chin-Tser H.; Matthews M., "Secure multicast in various scenarios of WirelessMAN", SoutheastCon, 2007. Proceedings. IEEE, Page(s):709 - 714, March 2007 [119]. Ginley M.; Sen X.; Chin-Tser H.; Matthews M., "Efficient and Secure Multicast in WirelessMAN", Wireless Pervasive Computing, 2007. ISWPC '07. 2nd International Symposium on, 5-7 Feb. 2007 [120]. Moiseev S.N.; Sergey N. M.; Filin S.A.; Kondakov M.S.; Alexandre V. G.; Garmonov A.V.; Andrew Y. S.; Savinkov A.Y.; Yun Sang P.; Seok Ho C., "Optimal Average Number of Data Block Transmissions for the ARQ Mechanism in the IEEE 802.16 OFDMA System", Personal, Indoor and Mobile Radio Communications, 2006 IEEE 17th International Symposium on, Page(s):1 - 5, Sept. 2006 [121]. Gowda H.; Lakshmaiah R.; Kaur M.; Mohanram C.; Singh M.; Dongre S., "A slot allocation mechanism for diverse QoS types in OFDMA based IEEE 802.16e systems", Advanced Communication Technology, The 9th International Conference on, Volume 1, Page(s):13 - 17, 12-14 Feb. 2007 [122]. Wang F.; Ghosh A.; Love R.; Stewart K.; Ratasuk R.; Bachu R.; Sun Y.; Zhao Q., "IEEE 802.16e system performance: analysis and simulations", Personal, Indoor and Mobile Radio Communications, 2005. PIMRC 2005. IEEE 16th International Symposium on, Volume 2, page(s): 900- 904, 11-14 Sept. 2005 [123]. Sen I.; Wang B.; Matolak David W., "Performance of IEEE 802.16 OFDMA Standard Systems in Airport Surface Area Channels", Integrated Communications, Navigation and Surveillance Conference, 2007. ICNS '07, Page(s):1 - 12, April 30 2007-May 3 2007 [124]. Hongxiang L.; Hui L., "An Analysis on Uplink OFDMA Optimality", Vehicular Technology Conference, 2006. VTC 2006-Spring. IEEE 63rd, Volume 3, Page(s):1339 - 1343, 2006 [125]. Esli C.; Delic H., "Performance Analysis for OFDMA in the Presence of Tone Interference", Communications, IEEE Transactions on, Volume 55, Issue 5, Page(s):845 - 849, May 2007 [126]. Kivanc D.; Hui L., "Subcarrier allocation and power control for OFDMA", Signals, Systems and Computers, 2000. Conference Record of the Thirty-Fourth Asilomar Conference on, Volume 1, Page(s):147 - 151, 29 Oct.-1 Nov. 2000 [127]. Hui-Juan Y.; Geng-Sheng K., "An Integrated QoS-Aware Mobility Architecture for Seamless Handover in IEEE 802.16e Mobile BWA Networks", Military Communications Conference, 2006. MILCOM 2006, Page(s):1 - 7, 23-25 Oct. 2006 [128]. Kuo G. S.; Yao H. J., "A QoS-Adaptive Admission Control for IEEE 802.16e-based Mobile BWA Networks", Consumer Communications and Networking Conference, 2007. CCNC 2007. 2007 4th IEEE, Page(s):833 - 837, Jan. 2007 [129]. Vatsa Omanand J.; Raj M.; K Ritesh K.; Panigrahy D.; Das D., "Adaptive Power Saving Algorithm for Mobile Subscriber Station in 802.16e", Communication Systems Software and Middleware, 2007. COMSWARE 2007. 2nd International Conference on, Page(s):1 - 7, 7-12 Jan. 2007 [130]. Zhang Y.; Fujise M., "Energy management in the IEEE 802.16e MAC", Communications Letters, IEEE, Volume 10, Issue 4, Page(s):311 - 313, Apr 2006 [131]. Xiao Y., "Energy saving mechanism in the IEEE 802.16e wireless MAN", Communications Letters, IEEE, Volume 9, Issue 7, Page(s):595 - 597, July 2005

107

[132]. Mukherjee S.; Leung K.K.; Rittenhouse G.E., "Protocol and control mechanisms to save terminal energy in IEEE 802.16 networks", Communications, Computers and signal Processing, 2005. PACRIM. 2005 IEEE Pacific Rim Conference on, Page(s):5 - 8, 24-26 Aug. 2005 [133]. Jee-young S.; Hyun-ho C.; Hyun-dae K.; Sang-wook K.; Dong-ho C.; Hong-sung C.; Geunwhi L.; Jun-hyung K., "Performance comparison of 802.16d OFDMA, TD-CDMA, cdma2000 1xEV-D0 and 802.11a WLAN on voice over IP service", Vehicular Technology Conference, 2005. VTC 2005-Spring. 2005 IEEE 61st, Volume 3, Page(s):1965 - 1969, 30 May-1 June 2005 [134]. Yuexing P.; Hongzhong Y.; Young il K.; Yong Su L.; Wenbo W., "Performance of Convolutional Turbo Coded High-speed Portable Internet (WiBro) System", Vehicular Technology Conference, 2007. VTC2007-Spring. IEEE 65th, Page(s):730 - 734, 22-25 April 2007 [135]. Taesoo K.; Howon L.; Sik C.; Juyeop K.; Dong-Ho C.; Sunghyun C.; Sangboh Y.; Won- Hyoung P.; Kiho K., "Design and implementation of a simulator based on a cross-layer protocol between MAC and PHY layers in a WiBro Compatible.IEEE 802.16e OFDMA system", Communications Magazine, IEEE, Volume 43, Issue 12, Page(s):136 - 146, Dec. 2005 [136]. Xiao Y.; Pan Y.; Li J., "Design and analysis of location management for 3G cellular networks", Parallel and Distributed Systems, IEEE Transactions on, Volume 15, Issue 4, Page(s):339 - 349, April 2004 [137]. Derryberry R.T.; Gray S.D.; Ionescu D.M.; Mandyam G.; Raghothaman B., "Transmit diversity in 3G CDMA systems", Communications Magazine, IEEE, Volume 40, Issue 4, page(s): 68-75, Apr 2002 [138]. Proctor T.K., "Evolution to 3G services: provision of 3G services over GERAN (GSM/EDGE radio access network)", 3G Mobile Communication Technologies, 2003. 3G 2003. 4th International Conference on (Conf. Publ. No. 494), Page(s):78 - 82, 25-27 June 2003 [139]. Rikkinen K.; Wildey C., "WCDMA scenarios for the 2.5 GHz IMT-2000 extension band supporting asymmetric frequency allocations", 3G Mobile Communication Technologies, 2003. 3G 2003. 4th International Conference on (Conf. Publ. No. 494), Page(s):294 - 298, 25-27 June 2003 [140]. Goransson B.; Hagerman B.; Petersson S.; Sorelius J., "Advanced antenna systems for WCDMA: link and system level results", Personal, Indoor and Mobile Radio Communications, 2000. PIMRC 2000. The 11th IEEE International Symposium on, Volume 1, Page(s):62 - 66, 18-21 Sept. 2000 [141]. Suryanegara M.; Hutabarat E.R.; Gunawan D., "The Interference on WCDMA System in 3G Coexistence Network", Personal, Indoor and Mobile Radio Communications, 2006 IEEE 17th International Symposium on, Page(s):1 - 5, Sept. 2006 [142]. Liangchi H.; Cheng M.W.; Niva I., "Evolution towards simultaneous high-speed packet data and voice services: an overview of cdma2000 1/spl times/EV-DV", Telecommunications, 2003. ICT 2003. 10th International Conference on, Volume 2, Page(s):1313 - 1317, 23 Feb.-1 March 2003 [143]. Becker G.E.; Rudrapatna R.; Sowlay S.; Wong K.N.; Wu J.R., "Integrated Network and Element Management System for the 3rd generation CDMA2000 wireless network",

108

Network Operations and Management Symposium, 2000. NOMS 2000. 2000 IEEE/IFIP, Page(s):953 - 954, 10-14 April 2000 [144]. Chen H. H.; Fan C. X.; Lu W. W., "China's perspectives on 3G mobile commuunications and beyond: TD-SCDMA technology", Wireless Communications, IEEE [see also IEEE Personal Communications], Volume 9, Issue 2, Page(s):48 - 59, April 2002 [145]. Xirrus High Performance Wi-Fi (http://www.xirrus.com/) [146]. Foundry Networks (http://www.foundrynet.com/) [147]. Interview From FSU-OTC, 10/08/2007, Mr. Clint Ringgold, Network Administrator of OTC and Mr. Phillip M. Callahan, Assistant Director of OTC. [148]. Interview From Q-WARE SYSTEMS & SERVICES CORP., 11/06/2007, Ms. Linda Yeh, Account Manager of Sales & Customer Service Dep. [149]. Tan K.; Feng J.; Qian Z.; Xuemin S., "Congestion Control in Multihop Wireless Networks", Vehicular Technology, IEEE Transactions on, Volume 56, Issue 2, March 2007 [150]. Yung Y.; Shakkottai, S., "Hop-by-hop Congestion Control over a Wireless Multi-hop Network", Networking, IEEE/ACM Transactions on, Volume 15, Issue 1, Feb. 2007 [151]. Zhou, H.; Hoang, D.; Nhan, P.; Mirchandani, V., "Introducing feedback congestion control to a network with IEEE 802.11 wireless LAN", Wireless Telecommunications Symposium, 2004, 14-15 May 2004 [152]. Weirong L.; Jianqiang Y.; Dongbin Z.; Wen, J.T., "A Smooth Control to Avoid Congestion for Wireless Network", Networking, Sensing and Control, 2006. ICNSC '06. Proceedings of the 2006 IEEE International Conference on, 23-25 April 2006 [153]. Seungcheon K.; Jamalipour, A., "Congestion control for best-effort services in wireless access network", ATM (ICATM 2001) and High Speed Intelligent Internet Symposium, 2001. Joint 4th IEEE International Conference on, 22-25 April 2001 [154]. IEEE 802.11u ( http://www.ieee802.org/11/Reports/tgu_update.htm)

109

BIOGRAPHICAL SKETCH Ming-Chieh Wu was born in Tainan, Taiwan, in 1978. He entered Tang Kong University in Taiwan, in 1998 majoring in Aerospace Engineering.. His interest at that time was navigation. In senior year, he joined a project called “Regional Aircraft Design” with another classmate. In this project, the group did not need to build a real airplane, but they had to run a simulation that had a reasonable solution to make sure this aircraft could fly. At the end of the project, the group received an “A.” Mr. Wu graduated in 2002 and joined the Air Force for two years. After his military service, he started to apply to graduate school in the U.S. and was accepted by Florida State University. In graduate school, he changed his major to Electrical Engineering. After his first semester, he found that his true interest was in wireless communication. After that, he started to take communication related classes. After graduating from FSU, Mr. Wu would like to design wireless communication networks because he believes that it will be the future trend.

110