214 Part 2: Planning and Designing Data Applications equalization and signal coherent combining are actually implemented jointly in the proposed scheme under a relatively simple hardware structure. 3. It operates adaptively to the channel characteristic variation without needing prior knowledge of the channel, such as interpath delay and relative strength of different paths. On the contrary, a RAKE receiver 0 ⌺ 0 ⌺ 0 ⌺ 0 0 0 ⌺ y(n)x(n) Shift register Initial 0 ⌺ 0 ⌺ 0 ⌺ 0 3 = 3 – 0 – 0 – 0 – 0 3 ⌺ y(n)x(n) Shift register Step 1 0 ⌺ 0 ⌺ 0 ⌺ 3 2 = 5 – 0 – 0 – 0 – 3 5 ⌺ y(n)x(n) Shift register Step 2 0 ⌺ 0 ⌺ 3 ⌺ 2 1 = 6 – 0 – 0 – 3 – 2 6 ⌺ y(n)x(n) Shift register Step 3 0 ⌺ 3 ⌺ 2 ⌺ 1 0 = 6 – 0 – 3 – 2 – 1 6 ⌺ y(n)x(n) Shift register Step 4 3 ⌺ 2 ⌺ 1 ⌺ 0 0 = 6 – 3 – 2 – 1 – 0 6 ⌺ y(n)x(n) Shift register Step 5 3 ⌺ 2 ⌺ 1 ⌺ 0 ⌺ y(n)x(n) Shift register Step 6 P 0 P 1 P 2 P 3 Send the multipath factor to multipath signal receiver Figure 7-13 A step-by-step illustra- tion of channel impulse response esti- mation using a recur- sive multipath signal reception filter. Chapter 7: Architecting Wireless Data Mobility Design 215 ⌺ ⌺ ⌺ T c T c T c sgn() 0 0 000 0 1 2 1 /3 Decision device y(n)x(n) Initial –––– ⌺ ⌺ ⌺ T c T c T c sgn() 0 3 000 1 = (3 – 0 – 0 – 0) ϫ 1 /3 1 2 1 /3 Decision device y(n)x(n) Step 1 –––– ⌺ ⌺ ⌺ T c T c T c sgn() 0 –1 000 –1 = (–1 – 0 – 0 – 2) ϫ 1 /3 1 2 1 /3 Decision device y(n)x(n) Step 2 –––– ⌺ ⌺ ⌺ T c T c T c sgn() 0 2 000 1 = (2 – 0 – 1 + 2) ϫ 1 /3 1 2 1 /3 Decision device y(n)x(n) Step 3 –––– ⌺ ⌺ ⌺ T c T c T c sgn() 0 –2 000 –1 = (–2 – 0 + 1 – 2) ϫ 1 /3 1 2 1 /3 Decision device y(n)x(n) Step 4 –––– ⌺ ⌺ ⌺ T c T c T c sgn() 0 2 000 1 = (2 – 0 – 1 + 2) ϫ 1 /3 1 2 1 /3 Decision device y(n)x(n) Step 5 –––– ⌺ ⌺ ⌺ T c T c T c sgn() 0 1 000 0 = (1 – 0 + 1 – 2) ϫ 1 /3 1 2 1 /3 Decision device y(n)x(n) Step 6 Step 7 –––– The binary stream is recovered y(n) = [1 –1 1 –1 1] Figure 7-14 The signal detection procedure of the recursive multipath signal reception filter based on channel impulse response estimates with recov- ered bit stream y(n) ϭ (1 Ϫ 11Ϫ 1 1). 216 Part 2: Planning and Designing Data Applications in a conventional CDMA system requires the path gain coefficients for maximal ratio combining, which themselves are usually unknown and thus have to be estimated by resorting to other complex algorithms. The performance of the proposed new CDMA architecture with the recursive filter for multipath signal reception is shown in Figs. 7-15 and 7-16, where two typical scenarios are considered: one for downlink per- formance and the other for uplink performance, similar to the perfor- mance comparison made for the MAI-AWGN channel in Figs. 7-8 and 7-9. 3 It is observed from the figures that, in terms of the BER in a synchro- nous downlink channel, three different codes perform similarly, whereas in an asynchronous uplink channel, the Gold code and m-sequence per- formances are much worse than the CC code, because the orthogonality among both Gold codes and m-sequences is destroyed by asynchronous bit streams from different mobiles. Nevertheless, the CC-code-based CDMA system outperforms conventional CDMA systems using either Gold code or m-sequence by a comfortable margin that can be as large as 4 to 6 dB, because of its superior MAI-independent property. Bandwidth Efficiency Previously in this chapter, it was demonstrated that the CDMA architec- ture based on CC codes and an adaptive recursive multipath signal reception filter is feasible and performs well. The system offers MAI-free M-seq 4-use RAKE Gold 4-use RAKE CCC 4-use recursive filter 10 –5 10 –4 10 –3 10 –2 10 –1 012345 E b /N 0 (dB) 678910 BER (for the first user) Figure 7-15 Downlink (synchro- nous) BER for CC- code-based CDMA and conventional CDMA systems in a multipath channel, with normalized mul- tipath power; inter- path delay ϭ 3 chips; multipath channel delay profile ϭ [1.35,1.08, 0.13]; PG ϭ 63/64; Gold code/m-sequence with MRC-RAKE; CC-code-based CDMA with the recur- sive filter. Chapter 7: Architecting Wireless Data Mobility Design operation for both down- and uplink transmissions in an MAI-AWGN channel. Another interesting property of the new CDMA system is its agility in changing the data transmission rate, which can be finished on the fly without needing to stop and search for a code with a specific spreading fac- tor, as required in the W-CDMA standards. Therefore, the rate-matching algorithm in the proposed system has been greatly simplified. Yet another important point that has to be addressed is the bandwidth efficiency of the proposed CDMA architecture. Spreading efficiency in bits per chip has been used to measure the bandwidth efficiency of a CDMA system because the bandwidth of a CDMA system is determined by the chip width of the spreading codes used. Table 7-3 compares the SEs of three systems: conventional CDMA and CC-code-based CDMA with and 217 PG 8 64 512 4096 32,768 262,144 Conventional CDMA 1/8 1/64 1/512 1/4096 1/32,768 1/262,144 CC-code-based CDMA 1/8 1/16 1/32 1/64 CC-code-based CDMA 1 1/8 1/16 1/32 (orthogonal carriers) TABLE 7-3 Spreading Efficiency (in Bits per Chip) Comparison of a Conventional CDMA System and a CC-Based CDMA System with and without Orthogonal Carriers M-seq 4-user use RAKE Gold 4-user use RAKE CCC 4-user use recursive filter 10 –5 10 –4 10 –3 10 –2 10 –1 01 23 4 5 E b /N 0 (dB) 67 8910 11 12 13 14 15 BER (for the first user) Figure 7-16 Uplink (asynchronous) BER for CC-code-based CDMA and conven- tional CDMA systems in a multipath chan- nel, with normalized multipath power; interpath delay ϭ 3 chips; interuser delay ϭ 2 chips; multipath channel delay profile ϭ [1.35,1.08, 0.13]; PG ϭ 63/64; Gold code/m-sequence with MRC-RAKE; CC-code-based CDMA with the recursive filter. 218 Part 2: Planning and Designing Data Applications without orthogonal carriers. 3 It is clear that the CC-code-based CDMA systems have a much higher SE figure than a conventional CDMA does, especially when the processing gain is relatively high. However, there exist some technical limitations for the proposed CC- code-based CDMA system, which ought to be properly addressed and can become the direction of possible future work for further improvement. Obviously, a CC-code-based CDMA system needs a multilevel digital modulation scheme to send its baseband information, because of the use of an offset-stacked spreading modulation technique, as shown in Figs. 7-6 and 7-7. If a long CC code is employed in the proposed CDMA system, the number of different levels generated from a baseband spreading modulator can be a problem. For instance, if the CC code of L ϭ 4 is used, as shown in Table 7-2, five possible levels will be generated from the offset-stacked spreading: 0, Ϫ2, and Ϫ4. However, if the CC code of L ϭ 16 in Table 7-2 is involved, the possible levels generated from the spreading modulator become 0, Ϫ2, Ϫ4, …, Ϫ16, comprising 17 different levels. In general, the modulator will yield L ϩ 1 different levels for a CC-code-based CDMA system using length L element codes. Given the element code length (L) of the CC code, it is necessary to choose a digital modem capable of transmitting L ϩ 1 different levels in a symbol dura- tion. An L ϩ 1 quadrature amplitude-modulated (QAM) digital modem can be a suitable choice for its robustness in detection efficiency. It should be pointed out that the simulation study concerned in this part of the chapter assumes an ideal modulation and demodulation process. Thus, the research takes into account the nonideal effect of multilevel carrier modulation, and demodulation remains a topic of future study. Finally, another concern with the CC-code-based CDMA system is that a relatively small number of users can be supported by a family of the CC codes. Take the L ϭ 64 CC code family as an example. It is seen from Table 7-3 that such a family has only eight flocks of codes, each of which can be assigned to one channel (for either pilot or data). If more users should be supported, long CC codes have to be used. On the other hand, the maximum length of the CC codes is in fact limited by the max- imal number of different baseband signal levels manageable in a digital modem, as mentioned earlier in this chapter. One possible solution to this problem is to introduce frequency divisions on top of the code divi- sions in each frequency band to create more transmission channels. Conclusion In this chapter, a new CDMA architecture based on CC codes was pre- sented, and its performance in both MAI-AWGN and multipath channels was evaluated by simulation. The proposed system possesses several Chapter 7: Architecting Wireless Data Mobility Design advantages over conventional CDMA systems currently available in 2G and 3G standards: 1. The system offers much higher bandwidth efficiency than is achiev- able in conventional CDMA systems. The system, under the same processing gain, can convey as much as 1 bit of information in each chip width, giving a spreading efficiency equal to 1. 2. It offers MAI-free operation in both synchronous and asynchronous MAI-AWGN channels, which attributes to cochannel interference reduction and capacity increase in a mobile cellular system. This excellent property also helps to improve the system performance in multipath channels, as shown by the obtained results. 3. The proposed system is inherently capable of delivering multirate/multimedia transmissions because of its offset-stacked spreading modulation technique. Rate matching in the new CDMA system becomes very easy, just shifting more or fewer chips between 2 consecutive bits to slow down or speed up the data rate—no more complex rate-matching algorithms. This chapter also proposed a novel recursive filter, particularly for multipath signal reception in the new CDMA system. The recursive fil- ter consists of two modules working jointly; one performing channel impulse response estimation and the other detecting signal contaminated by multipath interference. The recursive filter has a relatively simple hardware compared to a RAKE receiver in a conventional CDMA sys- tem, and performs very well in multipath channels. The chapter also addressed technical limitations of the new CDMA architecture, such as a relatively small family of CC codes and the need for complex multi- level digital modems. Nevertheless, the proposed CDMA architecture based on complete complementary codes offers a new option to imple- ment future wideband mobile communications beyond 3G. The increasing amount of roaming data users and broadband Internet services has created a strong demand for public high-speed IP access with sufficient roaming capability. Wireless data LAN systems offer high bandwidth but only modest IP roaming capability and global user man- agement features. This chapter described a system that efficiently integrates wireless data LAN access with the widely deployed GSM/GPRS roaming infra- structure. The designed architecture exploits GSM authentication, SIM- based user management, and billing mechanisms and combines them with public WDLAN access. With the presented solution, cellular operators can rapidly enter the growing broadband access market and utilize their existing subscriber management and roaming agreements. The OWDLAN system allows 219 220 Part 2: Planning and Designing Data Applications cellular subscribers to use the same SIM and user identity for WDLAN access. This gives the cellular operator a major competitive advantage over ISP operators, who have neither a large mobile customer base nor a cellular kind of roaming service. Finally, the designed architecture combines cellular authentication with native IP access. This can be considered the first step toward all-IP networks. The system proposes no changes to existing cellular network elements, which minimizes the standardization effort and enables rapid deployment. The reference system has been commercially implemented and successfully piloted by several mobile operators. The GSM SIM- based WDLAN authentication and accounting signaling has proved to be a robust and scalable approach that offers a very attractive opportu- nity for mobile operators to extend their mobility services to also cover indoor wireless data broadband access. References 1. “Wireless Architecture Options,” Synchrologic, 200 North Point Center East, Suite 600, Alpharetta, GA 30022, 2002. 2. “CIO Outlook 2001: Architecting Mobility,” Synchrologic, 200 North Point Center East, Suite 600, Alpharetta, GA 30022, 2002. 3. Hsiao-Hwa Chen, Jun-Feng Yeh, and Naoki Suehiro, “A Multicarrier CDMA Architecture Based on Orthogonal Complementary Codes for New Generations of Wideband Wireless Communications,” IEEE Com- munications Magazine, 445 Hoes Lane, Piscataway, NJ 08855, 2002. 4. John R. Vacca, The Essential Guide to Storage Area Networks, Prentice Hall, 2002. Fixed Wireless Data Network Design 8 CHAPTER 8 Copyright 2003 by The McGraw-Hill Companies, Inc. Click Here for Terms of Use. 222 Part 2: Planning and Designing Data Applications If you can’t wait for DSL or cable modem 3 to be installed at your corpo- rate headquarters or if it seems like broadband 4 will never be available at your remote sites, the design of a fixed wireless data network is becoming a viable alternative for last-mile Internet access. Fixed wireless data has some advantages over wired broadband: It can be installed in a matter of days. Once the line of sight is established, the connection isn’t susceptible to the types of weather-related or acci- dental outages that can occur with wired networks. But there are important design issues that network executives will need to resolve before signing up for fixed wireless data, including secu- rity and possible performance degradation from interference with other service providers. For example, on the island of Anguilla, a British territory 6 miles north of St. Martin in the Caribbean, Weblinks Limited (http://www.weblinksad- vertising.co.uk/contact_frameset.html) has installed a wireless data Inter- net system that covers the entire 16-mile-long island, offering services to a growing number of e-commerce 6 companies. On a hurricane-prone and remote island like Anguilla, fixed wireless data offers several benefits over DSL and cable modem. A fixed wireless Internet system, such as Weblinks’ in Anguilla, consists of centralized transceiver towers and directional antennas mounted at each end-user location to maximize range and mini- mize the number of towers needed to cover a large area (see sidebar, “Wire- less Data Internet Infrastructure”). (The Glossary defines many technical terms, abbreviations, and acronyms used in the book.) Wireless Data Internet Infrastructure Independent service providers are building private networks based on a combination of optical and fixed wireless data technology, exclusive peering arrangements, and Internet data centers to sup- port the B2B marketplace. The arrival of the twenty-first century in Latin America coincided with the migration of the region’s Inter- net from a communications/recreation medium to a platform for mission-critical applications and e-business. With this change, the region’s Internet infrastructure is evolving from its dependence on U.S based hosting facilities and incumbent owned and operated transport to a mix of fiber-optic and fixed wireless data private net- works with Internet data centers (IDCs). Until a few years ago, the dot-coms that pioneered Latin Ameri- can Web content looked to local garages or U.S based Web-hosting firms for their infrastructure needs, since high-quality solutions did not yet exist in the region. The distance between U.S. hosting Chapter 8: Fixed Wireless Data Network Design 223 facilities and Latin American users, combined with subpar infra- structure tying the two regions, resulted in poor performance and high-latency connections. Such concerns were not critical, however, because of the informational nature of the first Web sites. The ready-made U.S. solutions, which transported international traffic over satellite networks 5 or directed in-region traffic “hot-potato” style through multiple hubs and network access points (NAPs), suited both providers and users. Even today, many connections throughout the region suffer delay as a result of poor routing. For example, a user in Buenos Aires accessing a site hosted in California connects to an Internet service provider (ISP) that in turn connects to an Internet back- bone provider. Upon leaving the ISP network, the connection trav- els across the Internet “cloud.” The network providers inside the cloud have no incentive or ability to optimally route the connection. Their motivation is to minimize the costs by routing across inex- pensive and usually overly utilized links or by passing the session off to another less expensive and lower-quality network as soon as possible. This process, known as hot-potato routing, increases the number of hops and degrades the quality of the session. If a user connects to a local ISP in Argentina or Brazil to access content that is hosted in the same city or country, the user’s traffic is often routed to the United States, where it will be redirected at a public NAP back to its destination in South America. That occurs because of the limited partnerships at public access points and lack of peering agreements between local providers. The ISP’s backbone provider is likely an incumbent telecommu- nications provider with a legacy voice-based network. The legacy network’s routers and links can add significant latency and packet loss to the session. The provider’s network is also likely to include single points of failure that pose the risk of session failure. The precise number of hops, amount of packet loss, and amount of latency varies with each session and the network topologies of the connection. Generally, packets passing from sites in the United States to Buenos Aires would generate 500 ms or more of round-trip latency. Compounded by multiple packets making up a Web page, such latency can produce 8 s or more delay in page downloads. Today’s Pan-Regional Internet Backbone The Internet is entering the second phase of its evolution in Latin America. By 2000, the region emerged as the fastest-growing Inter- net market in the world. Companies no longer use the Web merely to market their products and services; many are developing highly [...]... (TDM) scheme is used Each RS can deliver voice and data using TABLE 8-1 The 3 .5- and 10 .5- GHz System Characteristics Product 3 .5 GHz 10 .5 GHz Frequency, GHz 3.4–3.6 10. 15 10. 65 Tx/Rx spacing, MHz 100 350 Channelization, MHz 3 .5, 5, 7 3 .5, 7 RS upstream modulation QPSK/16 QAM QPSK RS downstream modulation 16/64 QAM 16 QAM RS upstream capacity, Mbps 5 20 5, 10 RS downstream capacity, Mbps 12–34 12, 23 Coverage... Mina Danesh, Juan-Carlos Zuniga, and Fabio Concilio, “Fixed LowFrequency Broadband Wireless Access Radio Systems,” IEEE Communications Magazine, 4 45 Hoes Lane, Piscataway, NJ 08 855 , 2002 3 John R Vacca, The Cabling Handbook, 2d ed., Prentice Hall, 2001 4 John R Vacca, Wireless Broadband Networks Handbook, McGraw- Hill, 2001 5 John R Vacca, Satellite Encryption, Academic Press, 1999 6 John R Vacca, Electronic... Planning, Design, and Implementation, CRC Press, 2002 This page intentionally left blank 9 Wireless Data CHAPTER Access Design Copyright 2003 by The McGraw- Hill Companies, Inc Click Here for Terms of Use 238 Part 2: Planning and Designing Data Applications It is an exciting time for broadband5 fixed wireless data access design, with key developments in frequency bands from 1 to 60 GHz and a range of... Java.1 Now, let’s consider the ability of broadband fixed wireless data to play its envisaged role in this future vision Technical Constraints on Broadband Fixed Wireless Data Systems If fixed wireless data is to play a key role in this network of the future, it must be able to deliver high data rates to most homes Being more specific about high data rates is difficult because it depends on the user NOTE... Spectrum allocations above around 10 times the user data rate result in little extra increase in profitability; hence, a 10 times allocation is probably most appropriate, minimizing use of the scarce spectrum resource M (measure of radio channels per cell) 1.2 Figure 9-2 The relationship of M to SIR 1 0.8 0.6 0.4 0.2 0 5 10 15 20 25 30 SIR (dB) 35 40 45 50 2 45 Figure 9-3 Variation of profit with spectrum... very reasonable costs Fixed broadband wireless data access (BWDA) is a communication system that provides digital two-way voice, data, Internet, and video services, making use of a point-to-multipoint topology The BWDA low-frequency radio systems addressed in this part of the chapter are in the 3 .5- and 10 .5- GHz frequency bands The BWDA market targets wireless data multimedia services to small offices/home... assembled for horizontal or vertical polarization for reduced interference Conclusion Fixed wireless data is a good option for networks in locations where DSL and cable modem access are not available Small and midsize companies Chapter 8: Fixed Wireless Data Network Design 2 35 might also benefit from wireless data Internet in larger cities because of cost savings With the availability of the solid IEEE... 3G standard or, alternatively, multimode phones operating over a multiplicity of standards with high-speed access delivered by wireless data LAN (WD-LAN) solutions in certain important areas A standard approach for office wireless data networks, most likely based on wireless data LANs, common to all offices Communicator devices able to work on the cellular, home, and office networks in a seamless manner... 19 8 229 Edge router Figure 8-2 PSTN V5.2/GR.303 PSTN gateway STM-1/OC-3c STM-1/OC-3c Router and concentrator Radio tower Base station TDMA/TDM FDD 3 .5 GHz 10 .5 GHz Air interface Fixed broadband wireless data access system architecture CLEC ATM network Internet Network management and billing system IDU modem ODU radio Remote station E1/T1 clear channel E1/T1 V.35N ϫ 64 POTS 10/100 Base-T PBX Video LAN... multiple simultaneous transmissions to or from the home, data rates in excess of 10 Mbps will be needed To date, fixed wireless data has been unable to deliver data rates in excess of 10 Mbps to a high percentage of homes in a given area costeffectively In this part of the chapter, let’s examine the theoretical and economic constraints on fixed wireless data to assess whether this might change in the future . Each RS can deliver voice and data using Product 3 .5 GHz 10 .5 GHz Frequency, GHz 3.4–3.6 10. 15 10. 65 Tx/Rx spacing, MHz 100 350 Channelization, MHz 3 .5, 5, 7 3 .5, 7 RS upstream modulation QPSK/16. and improved intracountry networks, about 50 percent of the traffic in Chapter 8: Fixed Wireless Data Network Design Security Concerns Another key issue with wireless data Internet is security. A poorly secured. Prentice Hall, 2002. Fixed Wireless Data Network Design 8 CHAPTER 8 Copyright 2003 by The McGraw- Hill Companies, Inc. Click Here for Terms of Use. 222 Part 2: Planning and Designing Data Applications If