Deploying WiMAX Certified Broadband Wireless Access Systems

Transcription

1 Cristian Patachia-Sultanoiu Deploying WiMAX Certified Broadband Wireless Access Systems Author Cristian Patachia Sultanoiu Telecommunications Department of Electronics and Telecommunications Faculty, Politehnica University of Bucharest, B-du Iuliu Maniu nr. 1-3, Sector 6, Bucuresti, ROMANIA GSM: +40(0) , Tel: +40(0) , This paper presents the characteristics and requirements imposed for a future-proof broadband wireless access system deployment targeting the corporate, carrier and residential broadband wireless environment. Typical and particular WiMAX-based infrastructure scenarios are also discussed within the paper, including both the advantages and the constraints against the traditional technologies. Introduction Governments globally are starting to prioritise broadband as a key political objective for all citizens to overcome the broadband gap, also known as the divide gap. In first/last mile markets, where traditional cable or copper/fibre infrastructures are either saturated, outdated or simply out of reach, broadband wireless access (BWA) technology fills the void admirably, providing highly efficient and cost-effective access services for a large number of subscribers who would otherwise be left out of the loop in developed markets. According to Maravedis Inc. research, at the end of 2003 there were over pointto-multi-point (PtMP) BWA (sub 11 GHz) base stations (BSs) and 1.2 million customer premises equipments (CPEs) installed worldwide providing at least 256 kbit/s broadband services to over 1.5 million subscribers. EMEA, which represented 32% of the overall market in 2003, continues to represent the largest market opportunity. The carrier and private networks market segments represented respectively 85% and 15% of the total market in The access and backhaul applications represented respectively 84% and 16% of total sales in However, backhaul will represent 30% of equipment sales by GHz, the most allocated frequency band for BWA (Europe), represents the largest opportunity for BWA, representing 40% of total sales followed by the GHz band. Shipments of products based on orthogonal frequency-division multiplexing (OFDM) already represent 39% of all shipments and that proportion will grow with the mass adoption of a/RevD to close to 60% by Afterwards shipments of e will grow exponentially to 1 million units and will be dominated by Intel. IEEE Working Group Activities In 1998, many companies were beginning to develop and offer products for BWA, and spectrum was becoming available in many countries. The US National Institute of Standards and Technology (NIST) called a meeting to discuss the topic in August With the participation of 45 attendees, a plan was drafted to initiate the standardisation process. The group decided to organise within IEEE 802 and met with that organisation in November IEEE 802 accepted a proposal to initiate a Study Group activity to consider the problem. The Study Group s proposal, to standardise BWA from GHz, was accepted by IEEE, and IEEE Working Group (WG) held its first formal meeting in July In November 1999, while considering 35 technical proposals, the WG also initiated activity addressing 2 11 GHz. Holding bimonthly week-long sessions, the group moved quickly. The core standard defining the wireless metropolitan area network (WirelessMAN) air interface was approved as IEEE Standard and published in It addressed fixed line-ofsight (LOS) connections for the first mile/ last mile link and is focused on efficient use of various licensed frequencies in the GHz bandwidth c/d, published in January 2003, address interoperability by providing detailed system profiles and specifying combinations of options, as the basis for compliance and interoperability tests. The Worldwide Interoperability for Microwave Access (WiMAX) Forum presented the first of these tests in July 2003 and further work will be done by this body and the IEEE throughout this year. The c protocol relates to protocols, test suite structures and test purposes, while d fixes errata and protocols not covered in c, and creates the system profiles. The next version of the standard (802.16a), published in April 2003, is the 105

2 Figure 1 Typical WiMAX deployment scenarios Small/Large Enterprise Residential Best Connection for and nomadic users WAN Backbone (Usually MPLS based) Wi-Fi WISP Multi-Point Hot-Spot Backhaul Backhaul ISP one that has really kick-started WiMAX into being adopted as the dominant wireless broadband technology. This is also for fixed wireless but extends the range of WiMAX up to 50 km and operates in the lowfrequency 2 11 GHz spectrum and so can be adopted by unlicensed operators. It uses PtMP or (optionally) mesh topology and does not require LOS. Specifically, it uses licensed bands at 3.5 GHz and 10.5 GHz internationally and GHz in the US; and unlicensed 2.4 GHz and GHz. Currently, most of the work in the WG involves two main activities: 1 Project P REVd is a project to consolidate the three air interface standards (IEEE , a, and c), revising the content for technical enhancement and adding system profiles for the 2 11 GHz specification. The output, which should currently already be available, is a revised IEEE Standard incorporating the full specifications for 2 66 GHz fixed wireless access. 2 Project P802.16e is a project to amend IEEE , enhancing the air interface primarily by adding support for mobile user devices at vehicular speeds. The project will provide a BS specification that can support both fixed and mobile subscribers. The WG began addressing this subject in March 2002 and the outcome should be available at the end of Additional compliance test work is expected from both and WiMAX WGs during this year. IEEE WirelessMAN Specifications The IEEE a/RevD standard for 2 11 GHz is a wireless metropolitan area network (MAN) technology that will provide broadband wireless connectivity to fixed, portable and nomadic devices. Typically service providers will use the technology for the access networks. The s will be interconnected through the service provider wide area network (WAN), which in most of the cases is based on multiprotocol label switching (MPLS) technology. In this way the based quality-of-service (QoS) mechanisms will be enhanced by the MPLS QoS features allowing the service provider to guarantee the appropriate service level agreements (SLAs) offered to the customers. As Figure 1 shows, it can be used to connect hot spots to the, provide campus connectivity, and provide a wireless alternative to cable, digital subscriber line (DSL) and E1 for first mile/last mile broadband access. It provides up to 50 km of service area range, allows subscriber stations (SSs) to get broadband connectivity without needing direct line of sight with the base station (BS), and provides shared data rates (raw rates) of up to 70 Mbit/s per BS, which is enough bandwidth to simultaneously support hundreds of businesses with E1-type connectivity and hundreds of homes with DSL-like connectivity with a single BS. It specifies a protocol that among other things supports low latency sensitive applications such as voice and video, provides broadband connectivity without requiring a direct LOS between subscriber terminals and the BS and will support hundreds of subscribers from a single BS. The forthcoming e will add mobility to the standard and really throws down the gauntlet to cellular. This element of the standard has the particular interest of Nokia, which can see a new revenue stream at both BS and handset level. Because of the e draft delaying and not included by WiMAX into the first set of system profiles, Nokia has retired for the moment from WiMAX Forum. The draft was supposed to be ready in August 2003 but it is now postponed until the end of The standard (Table 1) operates at up to 124 Mbit/s in the 28 MHz channel (10 66 GHz), a/RevD at 70 Mbit/s in the lower-frequency 2 11 GHz spectrum. In reviewing the standard, the technical details and features that differentiate (and further WiMAX) certified equipment from certified Wi-Fi (Wireless Fidelity) or other technologies can best be illustrated by focusing on the two layers addressed in the standard, the physical (PHY) or radio frequency (RF) transmissions and the media access control (MAC) layer design PHY layer and the WiMAX PHY subset The IEEE standard was designed to evolve as a set of air interfaces based on a common MAC protocol but with PHY layer specifications dependent on the spectrum of use and the associated regulations. The MAC in general supports a PHY layer in which the BS basically transmits a time-division multiplex signal, with individual subscriber stations (SSs) allocated time-slots serially. Access in the uplink direction is by timedivision multiple access (TDMA). Both timedivision duplexing (TDD) and frequency-division duplexing (FDD) are handled in a common burst fashion. Halfduplex FDD SSs, which may be less expensive since they do not simultaneously transmit and receive, are easily supported in this framework, with a slight modification. Both TDD and FDD alternatives support adaptive burst profiles in which modulation and coding options are dynamically assigned on a burst-by-burst basis. 106

3 Table 1 IEEE specifications Criteria Initial a Amendment e Amendment Standard Completed December 2001 January 2003 Estimated end of 2004 Spectrum GHz 2 11 GHz Licensed bands 2 6 GHz Channel LOS only LOS, OLOS, NLOS LOS, OLOS, NLOS Conditions Raw bit Mbit/s 1 75 Mbit/s Up to 15 Mbit/s rate (28 MHz channel (20 MHz channel in 5 MHz channel bandwidth) bandwidth) bandwidth Modulation QPSK, 16QAM, 64QAM QPSK, 16QAM, 64QAM, OFDM 256 sub-carriers for single carriers (256QAM) for OFDM QPSK, 16QAM, 64QAM only 256 sub-carriers, OFDMA 2048 sub-carriers and single carrier UL/DL TDD/FDD TDD/FDD/HFDD TDD/FDD Duplexing Mobility Fixed Fixed, Portable Local/regional nomadic mobility (< 150 km/h), hand-off and roaming support Channel 20, 25 and 28 MHz Flexible channel Same as a with UL bandwidths between sub-channels 1.75 and 20 MHz depending on the regulatory bodies Typical cell 2 5 km 7 10 km, 2 5 km radius max range 50 km The GHz air interface is designed WirelessMAN-SC because it uses singlecarrier (SC) modulation. Channel sizes of 25 and 28 MHz are supported, with a maximum aggregate over-the-air data rate of 124 Mbit/s in the 28 MHz channel using 64- state quadrature amplitude modulation (64QAM). In typical cases, the operator will have access to tens of such channels, which will normally be used in an angular sector of perhaps 90. Range is typically of the order of less than 5 km. Compared to the higher frequencies, 2 11 GHz bands offer the opportunity to reach many more customers less expensively, though at generally lower data rates. This suggests that such services will be oriented toward individual homes or small and medium enterprises (SMEs). Design of the 2 11 GHz PHY layer was driven by the need for non-line-of-sight (NLOS) operation. This is essential to support residential applications, since rooftops may be too low for a clear LOS to a BS antenna, possibly due to obstruction by trees. Therefore, significant multipath propagation must be expected. Furthermore, outdoor-mounted antennas are expensive due to both hardware and installation costs. Some vendors will prefer to offer systems which use indoor antennas only. IEEE a/RevD specifies that systems implement one of three air interface specifications, each of which provides for interoperability: 1 WirelessMAN-SCa: using a single-carrier modulation format. 2 WirelessMAN-OFDM: using OFDM with a 256-point fast Fourier transform (FFT). Access is by TDMA. 3 WirelessMAN-OFDMA: This uses orthogonal frequency-division multiple access (OFDMA) with a 2048-point FFT. In this system, the multiple access is provided by addressing a subset of the multiple carriers to individual receivers. Although multiple PHYs are specified as in the suite of standards, the WiMAX Forum has determined that the first interoperable test plans and eventual certification will support the 256-point FFT OFDM PHY (which is common between a and ETSI HiperMAN), with the others to be developed as the market requires. The OFDM signalling format was selected in preference to competing formats such as code division multiple access (CDMA) due to its ability to support NLOS performance while maintaining a high level of spectral efficiency, maximising the use of available spectrum. In the case of CDMA (prevalent in 2G and 3G), the RF bandwidth must be much larger than the data throughput, in order to maintain processing gain adequate to overcome interference. This is clearly impractical for broadband wireless below 11 GHz, since, for example, data rates up to 70 Mbit/s would require RF bandwidth exceeding 200 MHz to deliver comparable processing gains and NLOS performance. Many systems in the past decade have involved fixed modulation, offering a tradeoff between higher-order modulation for high data rates, but requiring optimal links, or more robust lower order that will only operate at low data rates a/RevD supports adaptive modulation, balancing different data rates and link quality and adjusting the modulation method almost instantaneously for optimum data transfer and to make most efficient use of bandwidth. Some of the PHY layer features of a/RevD are included in Table MAC layer and the WiMAX MAC subset Every wireless network operates fundamentally in a shared medium and as such that requires a mechanism for controlling access by subscriber units to the medium. The MAC is quality of service (QoS)- sensitive and connection-oriented, with a BS Table a/RevD PHY system profiles elected by WiMAX Feature Benefit 256 point FFT OFDM waveform Built-in support for addressing multipath in outdoor LOS and NLOS environments. Adaptive modulation and variable error Ensures a robust RF link while maximising the correction encoding per RF burst number of bits/second for each subscriber unit. TDD, FDD and HFDD duplexing burst Address varying worldwide regulations where symmetric and asymmetric one or both may be allowed. Flexible channel sizes depending on the Provides the flexibility necessary to operate in licensed/le spectrum: 1.75 MHz, 3.5 MHz, many different frequency bands with varying 5 MHz, 7MHz, 10 MHz, 14 MHz, 20 MHz channel requirements around the world. Designed to support smart antenna systems Smart antennas are fast becoming more affordable, and as these costs come down their ability to suppress interference and increase system gain will become important to BWA deployments. 107

4 Figure basic DL sub-frame structure TDM Portion Tx/Rx Transition Gap (TDD only) SCh BCCh (QAM-4) TDM QAM-4 FEC a TDM QAM-16 FEC b TDM QAM-64 FEC c TDMA Portion Preamble QAM-x FEC d Preamble QAM-y FEC e Preamble QAM-z FEC 1 *** Preamble QAM-z FEC g Preamble DL Map PHY Control UL Map MAC Control Burst Start Points allocating bandwidth according to terminal requests. Access and bandwidth allocation algorithms accommodate hundreds of terminals per channel, with terminals that may be shared by multiple end-users. The request-grant mechanism is designed to maintain its efficiency when presented with multiple connections per terminal, multiple QoS levels per terminal, and a large number of statistically multiplexed users. It takes advantage of a wide variety of request mechanisms, balancing the stability of contentionless access with the efficiency of contention-oriented access. While extensive bandwidth allocation and QoS mechanisms are provided, the details of scheduling and reservation management are left unstandardised and provide an important mechanism for vendors to differentiate their equipment. The basic downlink sub-frame structure is shown in the Figure 2. The uplink is TDMA with each CPE bursting in an assigned slot using an individually assigned burst profile (modulation/forward error correction (FEC) combination). The frame structure allows terminals to be dynamically assigned uplink (UL) and downlink (DL) burst profiles according to their link conditions. This trade-off between capacity and robustness in real-time provides roughly a two times increase in capacity on average when compared to nonadaptive systems, while maintaining appropriate link availability. The MAC uses a self-correcting bandwidth request/grant scheme that eliminates the overhead and delay of acknowledgements, while simultaneously allowing better QoS handling than traditional acknowledged schemes (Wi-Fi). Terminals have a variety of options available to them for requesting bandwidth depending upon the QoS and traffic parameters of their services. They can be polled individually or in groups. They can steal bandwidth already allocated to make requests for more. They can signal the need to be polled, and they can piggyback requests for bandwidth. The MAC uses a variable-length protocol data unit (PDU) along with a number of other concepts that greatly increase the efficiency of the standard. Multiple MAC PDUs may be concatenated into a single burst to save PHY overhead. Additionally, multiple service data units (SDUs) for the same service may be concatenated into a single MAC PDU, saving on MAC header overhead. Fragmentation allows very large SDUs to be sent as pieces to guarantee the QoS of competing services. And payload header suppression can be used to reduce the overhead caused by the redundant portions of SDU headers. The MAC was designed specifically for PtMP wireless access environments. It is designed to seamlessly carry any higher-layer or transport protocol such as IPv4 and IPv6, packetised voice over protocol (VoIP), Ethernet or asynchronous transfer mode (ATM), and is designed to easily accommodate future protocols that have not yet been developed. The MAC is designed for the very high bit rates of the truly broadband PHY layer, while delivering ATM compatible QoS to ATM as well as non-atm (multiprotocol label switching (MPLS), VoIP, etc.) service. Along with the fundamental task of allocating bandwidth and transporting data, the MAC includes a privacy sublayer that provides reliable key exchange and encryption for data privacy using data encryption standard (DES) in CBC (chipper block chaining) mode with hooks defined for stronger algorithms like the advanced encryption standard (AES). Moreover, it provides authentication with x.509 certificates Table a MAC system profiles elected by WiMAX Feature TDM/TDMA scheduled uplink/downlink frames Scalable from one to hundreds of subscribers Connection-oriented QoS support: continuous grant real-time variable bit rate. non-real-time variable bit rate, best effort. Automatic retransmission request (ARQ) Support by adaptive modulation Security and encryption (triple DES) Automatic power control Benefit Efficient bandwidth usage. Allows cost-effective deployments by supporting enough subscribers to deliver a robust business case. Provide support for per-connection QoS and for faster packet routing and forwarding. Low latency for delay-sensitive services (TDM voice, VoIP), optimal transport for VBR traffic (video). Data prioritisation. Improves end-to-end performance by hiding RF layer induced errors from upper-layer protocols. Enables highest data rates allowed for channel conditions, improving system capacity. Protects user privacy. Enables cellular deployments by minimising self interference. 108

5 of network access and connection establishment to avoid theft of service. As can be seen, by the time authentication, security, capability negotiation and a host of other features are added, the IEEE standard becomes almost overwhelming. Table 3 gives a high-level overview of some of the MAC layer features of the IEEE a standard. IEEE and ETSI Joint Effort European standardisation work for fixed BWA has taken place under the auspices of the ETSI Broadband Radio Access Networks Projects HiperACCESS and HiperMAN. The HiperACCESS PtMP system is envisaged for licensed activity in frequency ranges above about 11 GHz and focuses on candidate frequency bands for fixed wireless access systems such as 26/28 GHz, 32 GHz and the 40 GHz multimedia wireless systems band. During early 2002, the first standards for HiperACCESS systems were approved and published. The IEEE a/RevD (256 FFT OFDM PHY) and ETSI HiperMAN standards share the same PHY and MAC. The WiMAX Forum is active in both standards making bodies to ensure that there is a global standard for wireless MAN. Moreover, the a WG and ETSI have exchanged over 50 liaison documents. It appears that decisions within the two bodies will ensure that the future ETSI HiperMAN specifications will be compliant with IEEE Standard a/RevD. WiMAX Approach It has been shown repeatedly that the adoption of a standard does not always lead to adoption by the intended market. The b wireless local area network (WLAN) standard was ratified in 1999; however, it did not reach mass adoption until the introduction of the Wi-Fi Alliance and certified interoperable equipment was available in For a market to be truly enabled, products must be certified that they do adhere to the standard first, and once certified it must also be shown that they interoperate. Interoperability means that end-users can buy the brand they like, with the features they want, and know it will work with all other like-certified products. The IEEE does not fulfil this role, leaving it to private industry to take a given technological standard and drive it that last crucial mile for mass adoption. In the case of WLANs, this role was and is fulfilled by the Wi-Fi Alliance. For the BWA market and its and ETSI HiperMAN standards, this role is played by the Worldwide Microwave Interoperability Forum or WiMAX. WiMAX is a non-profit industry trade organisation that has been chartered to remove an important barrier to adoption of the standard by assuring demonstrable interoperability between system components developed by OEMs. WiMAX will develop conformance and interoperability test plans, select certification labs and will host interoperability events (called Plug Fest) for equipment vendors. By defining and conducting interoperability testing, and by awarding vendor systems a WiMAX Certified TM label, WiMAX will model the approach pioneered by the Wi-Fi Alliance that ignited the WLAN industry, bringing the same benefits to the BWA market segment. Non-standardised equipment using proprietary interfaces is leading to high-risk one-vendor strategies. WiMAX plans to enforce standards compliance among vendor members. This compliance will result in interoperability and ultimately plug-and-play products, the cost of which will benefit from economics of scale and hence bring dramatic improvement to the business case for the operator. WiMAX members WiMAX was formed in April 2001 by Nokia (now retired), Ensemble and the OFDM Forum in anticipation of the publication of the original GHz IEEE specifications. Currently, there are more than 100 WiMAX member companies which include leading operators, equipment and component makers. In July 2003, Intel announced that it would develop silicon chips for WiMAX equipment. Wireless giant Nokia, chipmaker Fujitsu, and wireless equipment makers Alvarion and Proxim have also thrown their weight behind WiMAX; and start-ups like Aperto Networks and Ensemble Communications are also backing the standard. Other big names, including AT&T, Covad and Siemens, also joined the group this year. These companies hope to get a cut of what could be a sizeable market. The whole industry is benefiting from the late entry into the market of Intel which is behind most of the publicity around WiMAX. But Intel did not enter the game simply to address the fixed broadband wireless access (FBWA) market. It is betting heavily on the migration of chipsets into millions of mobile devices. WiMAX s products requirements have come from operators all over the world via the vendors who participate in the WiMAX Forum. The purpose of the WiMAX Forum is to ensure that all vendors create interoperable products. Operators have joined and are encouraged to participate in WiMAX and can become members. This allows direct participation in the Technical Working Groups (TWGs), and ensures that operator specific requirements are considered in the development of system profiles. According to WiMAX recommendations, operators who want to deploy today should deploy with equipment from today s WiMAX member companies. Current members of WiMAX account for over 75% of all sub 11 GHz BWA shipments worldwide, and in some regions this number is even higher. By deploying equipment from today s WiMAX members, an operator can be assured there is a roadmap and upgrade path to WiMAX-compliant equipment in the future. Currently there are commercially available some pre-wimax systems which are a compliant. There are significant names missing from WiMAX so far. Its initial focus on last/ first mile is indicated by the bias of the membership towards fixed (and portable and nomadic) wireless, rather than enterprise-focused suppliers or mobile carriers. But these companies, with proprietary solutions today, will join WiMAX Forum Cisco being a critical target. WiMAX working groups WIMAX created the GHz TWG. The profiles and test specifications are created by the TWG, but actual testing is done by an authorised independent laboratory. The WiMAX 2-11 GHz TWG has the mandate of creating testing and conformance documents as contributions to IEEE and ETSI standards bodies in support of the IEEE a/RevD and ETSI HiperMAN standards. A first set of documents should already be available. Similar to the first case, an authorised lab will conduct all the testing phases. Probably an important new project and new TWG, which could be considered by WiMAX members, will be to enable handoff between Wi-Fi and WiMAX. WiMAX roll-out In April 2003, WiMAX Forum selected the initial system profile (256 OFDM at 2.5 GHz, 3.5 GHz and 5.8 GHz) and started to complete the test suites. During 2004, WiMAX will develop conformance test plans, select certification labs and host interoperability events (Plug Fests) for a/RevD equipment vendors. The group is also currently working with the European Telecommunications Standards Institute (ETSI) to develop similar test plan for HiperMAN. Already some vendors are shipping a-compliant equipment. However, 109

6 first WiMAX certified products will be available by the beginning of The standard is set to revolutionise the BWA market with research showing that by 2008 up to 50% of all broadband wireless equipment could support this standard. Systems based on the mobile version of the standard, which should ship towards the end of 2005, about six months after fixed and portable wireless products, will be able to achieve long-distance wireless networking and will have far greater potential than Wi-Fi hot spots to provide ubiquitous coverage to rival that of the cellular network. WiMAX Testing Process and Interoperability Challenges From the above paragraphs, it is clear that the IEEE x Air Interface Specification is a very large specification. It was designed to cover the FBWA needs of a variety of different situations. Because of the wealth of options available, a system developer currently faces a tough decision. Does it build an IEEE compliant system implementing every possible feature, even those features it knows will never be used in systems for its target customers? Or, does it build a system with only the subset of features it needs for its market, risking accusations of non-compliance and lack of interoperability? WiMAX is focused on establishing a unique subset of baseline features grouped in what is referred to as system profiles that all compliant equipment must satisfy. The IEEE WG will then include these system profiles in the IEEE specification. These profiles and a suite of test protocols will establish a baseline interoperable protocol, allowing multiple vendors equipment to interoperate; the net result is that system integrators and service providers will have the option to purchase equipment from more than one supplier. System profiles For GHz, WiMAX currently is defining two primary PHY system profiles: single carrier with 25 MHz wide channel, which is typically used for US deployments; and single carrier with 28 MHz wide channel, which is typically used for European deployments. The PHY profiles are the same except for their channel width and their symbol rate, which is proportional to their channel width. Each primary PHY profile has two sub-profiles: TDD and FDD. Moreover for GHz, WiMAX is defining two MAC system profiles: a basic ATM system MAC profile, and a basic IP system MAC profile. The WiMAX 2 11 GHz TWG is currently defining MAC and PHY system profiles for IEEE a/RevD and HiperMAN standards. The MAC profiles that are being developed include IP-based versions for both WirelessMAN (licensed bands) and WirelessHUMAN (licence-exempt (LE) bands). While the IEEE a/RevD amendment has several PHY layer profiles, the WiMAX Forum through its 2 11 GHz TWG is focusing on the 256-point FFT OFDM PHY mode as its initial and primary interoperability mode. Various channel rasters covering typical spectrum allocations in both licensed and LE bands around the globe have been chosen, all supporting the 256-point FFT OFDM PHY mode of operation. This PHY will be combined with a non-optional MAC, ensuring a uniform base for all WiMAX implementations. Initially, there will be three system profiles, covering the 5.8 GHz LE band and the 2.5 GHz (multichannel multipoint distribution service (MMDS)) and 3.5 GHz (international fixed wireless access (FWA)) licensed bands. More system profiles are planned including the 2.3 GHz and 2.4 GHz bands and more. Test specifications Another issue facing IEEE developers is the fact that the IEEE standards process is concentrating primarily on requirements. The output of the IEEE WG is a standard, which means a requirement specification. The WG will continue to expand the standard to cover additional markets. This continuing work will result in amendments to the standard, but they will still address requirements. Initially, there was no work item in IEEE to address the creation of test specifications. Test specifications are necessary: to ensure that equipment and systems claiming compliance to the standard or a profile have been sufficiently tested to demonstrate that compliance; to guarantee that equipment from multiple vendors has been tested the same way, to the same interpretation of the standard, increasing the interoperability of the equipment; and to enable independent conformance testing, giving further credibility to the previous two items. WiMAX is establishing a structured compliance procedure based upon the proven test methodology specified by ISO/ IEC The process starts with standardised test purposes (TPs) written in English, which are then translated into standardised abstract test suites in a language called TTCN (Tree and Tabular Combined Notation). In parallel with the test purposes, the TPs are also used as input to generate test tables referred to as the PICS (Protocol Implementation Conformance Statement) Proforma. The end result is a complete set of test tools that WiMAX will make available to equipment developers so they can design in conformance and interoperability during the earliest possible phase of product development. Typically, this activity will commence when the first integrated prototype becomes available. Conformance statements A final issue facing developers of IEEE compliant systems is that having profiles is only part of the interoperability challenge. There must be a standard method of identifying which profiles a device or system complies with and which optional features are implemented so that system integrators can make educated decisions about specific features to provide to customers and to aid in the selection of equipment. On 25 March 2004, the IEEE-SA Standards Board approved a Project Authorisation Request (PAR), to develop a new IEEE conformance standard. The PAR is entitled Draft Standard for Conformance to IEEE Part 4: Protocol Implementation Conformance Statement (PICS) Proforma for Frequency bellow 11 GHz. This standard will include the references for BSs and SSs conformance specifications considering the air interface specified in IEEE RevD for frequencies below 11 GHz. The s Task Group C has been assigned to develop the draft until July When finished this document has to be completed by the supplier of a product claiming to implement the protocol. It will indicate which capabilities and options have been implemented. It will allow a user of the product to evaluate its conformance and to determine whether the product meets the user s requirements. For each system profile, functions are separated between mandatory and optional feature classes by the PICS Proforma document. There can be differences from one equipment manufacturer to another in implementing optional features, but mandatory features will be the same in every vendor s product. Implementation of an optional feature is noted when a vendor fills out the PICS Proforma. Spectrum Policy Challenges for WiMAX While all the features listed in Table 2 are necessary requirements for basic outdoor BWA operation, flexible channel sizes are required if a standard is to truly address worldwide deployment. This is because the regulations governing what frequency 110

7 equipment can operate in, and as a result the size of the channels used, can vary country by country. In the case of licensed spectrum where an operator had to pay for every MHz granted, it is imperative that the systems deployed use all the allocated spectrum and provide flexibility in either sectored or single sectored deployments. For example, a service provider in Europe operating in the 3.5 GHz FWA band, who has been allocated 14 MHz of spectrum, is likely to want equipment that supports 3.5 MHz and/or 7 MHz channel bandwidths (better if it also supports 1.75 MHz) for maximum flexibility and, depending on regulatory requirements, TDD or FDD operation. Therefore, it does not want a system that has 6 MHz channels, wasting 2 MHz of spectrum. Similarly, a wireless service provider (WISP) using LE spectrum in the 5.8 GHz band might desire equipment that supports TDD and a 10 MHz channel. Profiles can address these regulatory spectrum constraints faced by operators in different geographies. ETSI guidelines Although arguably not fully addressing all aspects of coexistence, voluntary standards have been developed by ETSI that cover (not exclusively) FBWA systems operating in specific frequency bands. These form a basis for inter-system coexistence and generally the basis for licensing approval and a route for demonstrating compliance with the Radio & Telecommunications Terminal Equipment (RTTE) Directive within the EU. Harmonised standard EN identifies the appropriate characteristics detailed in the numerous equipment and antenna standards that are considered essential requirements under the terms of the R&TTE Directive. The IEEE has also tackled FBWA coexistence issues and published a series of recommendations that, if followed, can promote equitable coexistence between systems conforming to their air interface standard. All systems that will be manufactured and/or deployed in Europe have to follow the recommendations specified by the following main standard bodies: European Radio Organization (ERO), European Telecommunications Standards Institute (ETSI) and International Organization for Standardization (ISO). National agency regulations rules ERO recommendations should be used as reference documents by the CEPT member countries when preparing their national regulations in order to keep in line with the provisions of the R&TTE Directive. In using these recommendations it should be remember that they represent the most widely accepted positions within the CEPT but it should not be assumed that all allocations are available in all countries. It should also be remembered that the pattern of radio use is not static. It is continuously evolving to reflect the many changes that are taking place in the radio environment, particularly in the field of technology. Spectrum allocation must reflect these changes and the position set out in the ERO recommendations is therefore subject to continuous reviews. Moreover, many administrations still have national allocations that do not conform to the CEPT positions. For these reasons, those wishing to develop, market and deploy WiMAX certified systems based on the European ERO RECs are advised to contact the relevant National Agency to verify that the positions set out in recommendations still apply. Some of the NAs are: UK s Ofcom (formerly Radiocommunications Agency), Germany s Regulative Authority, France s Autorite de Regulation des Telecommunications, etc. Licensed and licence-exempt bands policy Spectrum available either now or in the future, whether on a licensed or unlicensed basis, should be in harmony with allocations in other parts of the globe. Recent developments have greatly increased the availability of unlicensed spectrum but in isolation this may not be sufficient. LE spectrum minimises the entry barriers for potential operators but to some extent this is offset by the lack of protection from interference. Therefore the availability of licence-exempt spectrum should be balanced with licensed spectrum providing a migratory path for greater protection through exclusive assignments. This might be specifically true for longer-range systems like those standardised in More spectrum for licensed services would be beneficial. The standardisation of systems operating in licensed spectrum is a major element of the standard. For an operator wanting to provide a complete basket of services, operation under the two licensing conditions may be helpful. Operation within either LE or licensed spectrum might allow for a variety of service grades helping to encourage competitive offerings addressing differing sectors of the market. The standard provides for fully scheduled traffic to provide close control over the grade of service. It might be argued that licensed spectrum is more consistent with this feature. Moreover, LE transmissions are still subject to rules and constraints, such as power limit (TPC), dynamic frequency selection, etc. There are several different regulatory approaches that determine access to the spectrum for wireless broadband service providers. Service providers using networks composed of unlicensed devices do not pay for access to the spectrum, but must not cause interference and must share the spectrum with other operators of unlicensed devices, whereas access to other spectrum is obtained through licensing after successful bidding at auction. In addition, some spectrum has been made available on a first-come first-served basis. The applications of auction procedures have lead to examples around the world where legal considerations and obligations have carried greater weight than common-sense spectrum management. This can lead to nonoptimal or delayed spectrum access for new operators (or new technology) that is difficult to resolve in a timely manner. Alternatively, first-come first-served procedures can also lead to spectrum locked up in a way that similarly restricts the access. For wireless broadband to be successful, potential operators must have access to spectrum when they need it and in a way that is consistent with a growing network. The regulatory framework should provide a perception of ready access to spectrum of the appropriate quality for wireless broadband services to support the standardisation efforts of participants. The spectrum (and licensing framework) should be seen in the context of an overall spectrum allocation strategy that properly accounts for the potential for future growth in terms of services and user demand development. Deployment Scenarios Some of the main BWA equipment providers have completed a techno-economic assessment together to analyse the effect of introducing WiMAX-based equipment in various scenarios by performing detailed business-case analysis. The market investigated consists of both residential and business users who subscribe to various services similar to those offered by existing wired broadband operators. The BS equipment used to address these customers is modular and scalable, allowing operators to pay-as-they-grow. The most common a/RevD configuration consists of a BS mounted on a building or tower that communicates on a PtMP basis with subscriber stations (SS) located in business and homes (see Figure 1). Depending on topography and antenna height, IEEE a/RevD based equipment may achieve up to 50 km of range with a typical cell radius of 6 10 km. 111

8 Figure 3 WiMAX deployment scenarios for business PSTN PLMN N 64k Level Services for Small Enterprise WAN MPLS-Based Backbone with appropriate QoS NLOS PtP Backhaul VPN IP Services for SOHO N E1 Level Services for Large Enterprise VPN IP and VoIP Services for SME Since the high cost of CPEs has been the principal obstacle to making the customer fixed wireless business model work in the past, the business case has been considered for different types of CPE (depending on end-user needs): a modem attached to an external antenna; a modem with an indoor antenna; and an integrated antenna since, as further integration into silicon by major chip suppliers takes hold, CPEs can be integrated into laptops, phones and other devices (for example, PCMCIA (Personal Computer Memory Card International Association). The results of the investigation show that there is a positive business case for operators who want to add services and applications which are comparable to other existing broadband technologies (for example, cable or DSL) for both high-volume residential and high-revenue business customers in greenfield and overlay scenarios and to address the problems associated with limited range and hence limited penetration in underserved areas. Usage models for business users The MAC relies on a grant/request protocol for access to the medium and it supports differentiated service levels. Therefore, it could support in the same time dedicated E1 (G.703/G.704) circuits for business (Figure 3) and best-effort traffic for residential (Figure 5). The protocol employs TDM data streams on the DL and TDMA on the UL, with the hooks for a centralised scheduler to support delay-sensitive services like voice and video. By assuring collision-free data access to the channel, the MAC improves total system throughput and bandwidth efficiency, in comparison with contention-based access techniques like the carrier sense multiple access/collision avoidance (CSMA/CA) protocol used in WLANs. DSL typically operates at 128 kbit/s to 1.5 Mbit/s and slower on the upstream. Enterprise can use WiMAX (N E1) instead of traditional technologies for about 50% of the cost, while SMEs can be offered fractional E1 services (N 64 kbit/s). Supported access services are: access, virtual private network (VPN) IP, transparent LAN service (virtual LAN (VLANs) trunking), voice-over-ip transport and TDM transport. According to The Yankee Group, SME businesses will likely be a sweet spot for fixed wireless providers. While almost 90% of large enterprises have broadband access, only 35% of SME businesses do. But building a profitable business in this market will be no easy ride. Small businesses often look for the cheapest and easiest option. Companies will have little time for broadband if it requires a quasi-degree in wireless technology. Usage models for hot-spot backhauling Wi-Fi hot-spot operators may be able to build a spot for a few thousand dollars worth of equipment, but then they need to backhaul it to the public network and this is normally done with expensive E1 and DSL. WiMAX backhaul (Figure 4) could significantly reduce hot-spot costs, although there Figure 4 WiMAX deployment scenarios for Wi-Fi hot-spot backhaul NLOS PtP School Hot-Spot Wi-Fi University Campus Hot-Spot Wi-Fi Coffee Hot-Spot WAN Backbone Wi-Fi Wi-Fi Enterprise Hot-Spot Wi-Fi Airport Hot-Spot Hotel Hot-Spot 112

9 Figure 5 WiMAX deployment scenarios for residential users Integrated Indoor s Integrated Outdoor High-Speed Access for Residential Users WAN Backbone Ethernet Cable Remote VPN Service for Residential/Business Users Integrated Indoor & SS Home Yard Residential WDSL and Home Networking Residential WDSL and Home Networking is also the potential for Wi-Fi to be bypassed altogether by WiMAX hotzones. Usage models for residential users Regarding the system specifications of the current a/RevD compliant solutions and future WiMAX certified products, there are different flavours and colours that definitively would make a price/performance comparison very difficult. Some systems only offer data services at Ethernet 10/100 BaseT port level including VoIP with additional VoIP gateways; others also provide E1 (G.703/G.704) and POTS (plain old telephone system) based on analogue voice ports. Some systems are plug and play; others are outdoor only and some are indoor only. Some offer 2 Mbit/s at the CPE while others, 512 kbit/s, etc. Figure 5 shows the most important WiMAX deployment architectures for residential users considering the above issues and the most common residential services. Many in the industry say that the success of fixed wireless hinges on access providers securing a steady stream of revenue from business customers, before tackling residential markets. It is like the cellular industry, where cell-phone companies went after business subscribers first. The same thing will happen in this market. For the residential market, the business case may not be there yet. Usage models for WISP backhauling WISPs will likely set out pizza-box-sized antennas to homes and businesses. Ideally, customers install their own antennas, eliminating the cost of sending out an engineer. The WISP then beams a signal to dozens of businesses or hundreds of homes from a single base station. (See Figure 6.) Usage models for nomadic and portable applications The greatest media excitement about WiMAX has centered on its potential mobility and its role as a backhaul or even replacement for public Wi-Fi through its portability and nomadicity features (Figure 7). However, its initial raison d etre, and still its primary focus, is on FBWA for homes and businesses. In 2006, we will see the start of the second stage in the WiMAX revolution with WiMAX chipsets embedded in laptops and other mobile devices. Classical LOS versus NLOS challenge The a/RevD standard is designed for optimal performance in all types of propagation environments, including LOS, near LOS or obstructed LOS and NLOS environments, and delivers reliable robust performance even in cases where extreme link particularities have been introduced. The robust OFDM waveform supports high spectral efficiency (bits per second per Hertz) over ranges from 2 50 km with up to 70 Mbit/s in a single RF channel. Advanced topologies like mesh networks and antenna techniques like beam-forming, space time coding (STC) and antenna diversity can be employed to improve coverage even further. These advanced techniques can also be used to increase spectral efficiency, capacity, reuse, Figure 6 WiMAX deployment scenarios for WISP backhauling Carrier WAN Backbone WISP 2 Multi-Point Backhaul WISP 1 ISP Figure 7 WiMAX deployment scenarios for nomadic and portable users Laptop Connected through PCMCIA for Nomadic Users WAN Backbone (Usually MPLS Based) Best Connection for and Portable Users AP 113

10 and average and peak throughput per RF channel. Smart antenna mechanisms are one of the most important methods of improving spectral efficiency in non-cellular wireless networks and an important technique for NLOS cellular environments standards allow vendors to support a variety of these mechanisms, which can be a key performance differentiator. Therefore, mesh mode is a second optional topology for NLOS communication in a/RevD. The single-carrier PHY specification could represent an option for vendors that think they can beat multipath problems in this mode. OFDM will almost certainly become dominant in all wireless technologies including cellular, and its industry body, the OFDM Forum, is a founder member of WiMAX Forum. For NLOS operation, OFDM-based systems offer better multipath protection (due to intersymbol interference) than CDMA-based x systems. The specification is built on an implementation of OFDM from Wi-LAN. The above mentioned technical advances that will be incorporated in the future WiMAX-compliant equipments will result in a path loss performance expected to provide indoor coverage in at least 95% of all building within a 1 2 km radius from any base station in a cellular network. Key to coverage and NLOS capabilities is the concept of path loss. Path loss is a measure of the impact of how radio signals are affected by distance between BS and the SS, interaction with clutter in the area around the SS (such thing as building, trees and other objects) and the loss of signal strength that happens when penetrating through walls or leaking into a building where a SS may be located. Path loss is usually measured in decibels (db) which is a logarithmic (rather than a straight line) measure of the ratio of a signal strength measured at the BS compared to the signal strength measured where the SS is located. Therefore, a technology that can support a higher measure of path loss provides better performance than a technology that supports a lower measure of path loss. And this technology is supposed to be WiMAX. To achieve the anticipated indoor coverage at a radius of 1 2 km, a path loss capability in excess of 160 db is required, supposing 3.5 GHz band. Moreover, the farther up the spectrum chart a signal is generated, the greater its bandwidth potential. As frequencies drop to longer wavelengths, penetration increases. Therefore a 2.4 GHz radio will penetrate leaves and walls better than a 3.5 GHz or 5.8 GHz radio. However, it will inherently deliver less bandwidth. Alternative Technologies WiMAX competes directly with local phone and cable companies wired offerings, such as digital subscriber line (DSL), cable modems and leased lines, which currently dominate the broadband market. Another backhaul option that has gained coverage recently is free-space optical, a technology that is primarily used to extend local fibre networks but can also be used for backhaul. Moreover, satellite is another alternative technology. Still early in its lifecycle and potentially a powerful technology to integrate with WiMAX satellite has severe limitations of upstream bandwidth and spectrum availability, and suffers from high latency. There are some niche players with expensive equipment using proprietary mechanisms like smart antennas, OFDM and mesh to provide often effective alternatives to wired communications. A more realistic approach is that both proprietary systems and WiMAX are aiming at improving the coverage and penetration limitations of existing systems. The fact is that no system can go beyond the laws of physics and every deployment will face different challenges. There is no doubt that IP-based technologies cope with many of 3G s limitations in terms of data transfer, and provide a lower-cost installation of local loop, which is not dependent on LOS transmission as most early solutions were. But they are still weak on voice VoIP has quality issues and requires specialised phones. In less than five years, though, we should have VoIP cellphones that will finally establish IP as the dominant technology for mobile broadband communications, sidelining the descendants of both 3G and x. Currently, single-carrier PHY layer dominates deployments with about 50% of modems shipped worldwide. However, Maravedis Inc. forecasts both single-carrier and CDMA systems will loose market share, as OFDM-based a/RevD becomes the widely adopted standard for 2 11 GHz air interface of BWA systems. Various flavours of CDMA should however remain strong in niche markets for mobility. It is however difficult to sub-segment the whole market on system capabilities. Therefore, it is better and also recommended to keep in mind that a standard approach is always the secure way to be followed in open markets versus The WiMAX Forum is keen to present as complementary to the In many ways this is right. But there is conflict too. WiMAX makes redundant the efforts of Wi-Fi specialists to extend the reach of their favourite technology and places into a far smaller role than its supporters have looked for it. For years, the wildly successful or Wi-Fi technology has been used in BWA applications along with a number of proprietary-based solutions. When the WLAN technology was examined closely, it was evident that the overall design and features set available were not well suited for outdoor BWA applications. It could be done, it is being done, but with limited capacity in terms of bandwidth and subscribers, range and a host of other issues; it is clear this approach, while a great fit for indoor WLAN, was a poor fit for outdoor BWA. This is the opportunity for wireless technologies finally to grow up and offer a new, more complex and fully developed standard for addressing both the PHY layer environment (outdoor versus indoor RF transmissions) and the MAC-based quality of service needs demanded by the BWA and first/last mile access market. The most obvious differences between the two IEEE standards are stated in Table 4. The most constructive approach is that Wi-Fi and WiMAX are strongest when working together however. Some mobile operators are looking at offering a single PCMCIA card for roaming between and broadband services, but the big device makers will be developing cards for Wi-Fi/ WiMAX, and the debut of an Intel roaming card in the Centrino range will revolutionise the roaming hot-spot user s experience versus The big names are gathering behind two IEEE standards that, for all their claims of being complementary, are heading for a collision that involves more than wrestling in technical committees, but could instead be the cover for a serious battle for influence over the evolving wireless market. On the one hand we have WiMAX, on the other, , nicknamed Mobile-Fi (Mobile Fidelity), the first standard to be specifically designed from the outset to carry native IP traffic for fully mobile broadband access. It will provide symmetrical wireless rates from 1 Mbit/s to 4 Mbit/s in licensed spectrum below 3.5 GHz over distances of about 15 km. This makes it lower powered than WiMAX but more intrinsically mobile, offering latency of 10 ms even in a fast-moving vehicle, compared to 500 ms for 3G. A brief overview towards the technical differences between the two IEEE standard drafts is given in Table 5. Think of as providing services very similar to 3G, but with a data-centric, not voice-centric focus. 114

11 Table 4 Relationship between IEEE and IEEE Criteria Technical Difference Range Optimised for several Optimised for up to 50 km PHY tolerates greater multipath, delay spread hundreds of metres (add APs Typical cell size of 7 10 km. (reflections) via implementation of a 256 FFT versus for greater coverage). 64 FFT for Coverage Optimised for indoor Outdoor NLOS performance; systems have an overall higher system gain, performance, short range. standard support for advanced delivering greater penetration through obstacles antenna techniques. at longer distances. Scalability Intended for LAN applications, Designed to efficiently support The MAC protocol used in uses a CSMA/CA protocol, users scale from one to tens from one to hundreds of while employs a dynamic TDMA can only be with one subscriber for each CPEs, with unlimited used in licence-exempt spectrum with limited number of CPE device. subscribers behind each CPE. channels can use all available frequencies, multiple channels support cellular. Raw bit rate 2.7 bit/s/hz peak. 5 bit/s/hz peak. Higher modulations coupled with flexible error correction Up to 54 Mbit/s in 20MHz Up to 75 Mbit/s in a 20 MHz results in more efficient use of spectrum. channel. channel. QoS No QoS support into the QoS built into MAC for voice, is a contention-based MAC (CSMA/CA DCF with current standard e video and differentiated optional RTS/CTS and PCF) basically wireless Ethernet. will include EDCF/HCF service levels is a dynamic TDMA-based MAC with on-demand enhancements. bandwidth allocation. Table 5 Relationship between IEEE and IEEE Criteria e: a/RevD : proposed Mobile enhancement for Broadband Wireless Access Vehicular Mobility standard (Mobile-Fi) Completion Date End of 2004 End of 2004 Vehicular Mobility km/h Up to 250 km/h Spectrum Licensed bands GHz Licensed bands up to 3.5 GHz Channelisation Same as e with ability 1.25 MHz and 5 MHz to subchannelise into 8 or 16 channels in the UL (to support low-power devices at distance) Mobility Features Local/regional mobility Global mobility, hand-off and hand-off and roaming support roaming support Spectral Efficiency Up to 3 bit/s/hz Up to 3.2 bit/s/hz Peak Data Rate DL 70 Mbit/s for a 14 MHz channel 16 Mbit/s for a 5 MHz channel Peak Data Rate UL 70 Mbit/s for a 14 MHz channel 3.2 Mbit/s for a 1.25 MHz channel But has three critical weaknesses: WiMAX is starting to take on some of its initial supporters like Navini; WiMAX has stronger and more aggressive support from key vendors; and the mobile operators, while relatively friendly towards , are hostile to In addition, is still in the very early stages of standards development and is not expected to be completed before the end of 2004 with interoperability several years beyond that. WiMAX s place within 3G and 4G technologies A typical use of WiMAX equipment (802.16a/RevD case) in mobile operators networks would be BTS/node B to BSC/RNC wireless connections. However, standardisation through the IEEE x specifications raises the potential to provide an alternative or complements to 3G. Therefore the equipment providers, services providers and market analysts are split in two groups. The first one is considering WiMAX as an alternative technology to wideband code division multiple access (WCDMA) and/or CDMA The second group proposes a dual approach where WiMAX is seen as a complementary technology supposed to enhance and speed up the 3G worldwide deployments. According to the WiMAX member statements, IEEE e is not intended to compete with 3G or other truly mobile efforts. Even though, many are seeing this addition feature as a pure 4G functionality. However, work on this standard has just begun and will take some time to complete. The IEEE e extension will enable nomadic capabilities for laptops and other mobile devices, allowing users to benefit from metro area portability of an xdsl-like service (also called wireless DSL). This extension will boost development of built-in chipsets thus eliminating the external modem altogether and allowing transmission directly to the laptop. This built-in CPE could lead to a CPE-less business model, which makes the case even more compelling for an operator, because the user is subsidising the model. Because the initial applications can be limited to fixed users, an operator need not implement a geographically large deployment, with roaming, to get into business. As volume production brings down the cost, and the physical size, of the equipment, new applications may develop that take advantage of the standard s special features. After the e project is completed, the industry experience in building WirelessMAN SSs may allow for the rapid introduction of costeffective compliant mobile radio interfaces. At that time, existing deployments may be upgraded to support mobile terminals. The US FCC is freeing up more spectrum for metropolitan wireless networks. Such moves threaten the asset value of the 3G carrier s spectrum licences, since potentially competitive services can now be run over unlicensed bands. Europe is acting more slowly, but all territories will gradually take a similar direction and free up larger amounts of unlicensed spectrum, sacrificing licensing revenue for the government to the expected stimulus to business and the economy of better mobile communications. In Europe, progress is slower because the carriers have a more ubiquitous network and a vast financial investment in conventional cellular networks, and regulators have been less forward thinking. WiMAX is a serious threat to 3G because of its broadband capabilities, distance capabilities and ability to support voice effectively with full QoS. This makes it an alternative to cellular in a way that Wi-Fi can never be, so that while operators are integrating Wi-Fi into their offerings, looking 115

12 Figure 8 WiMAX for UMTS node B backhaul Bibliography UMTS Node B UMTS Node B to control both the licensed spectrum and the LE hot spots, they will have more problems accommodating WiMAX. But as with Wi-Fi, it will be better for them to take this path also than let independents do it for them, especially as economics and performance demands force them to incorporate IP into their systems. A first step for them, as can be seen in Figure 8, would be to use the first available WiMAX access networks for interconnecting the UMTS Node Bs with the radio network controllers (RNCs), decreasing in this way the CAPEX allocated to UMTS deployment. Handset makers such as Nokia will be working on this as they develop smartphones that support WiMAX as well as 3G. While WiMAX, like Wi-Fi, can be seen as a 3G alternative, it also offers opportunities to mobile carriers, to get into the last/first mile market and to build up their own hot-spot networks as an integrated service with 3G, the direction that most operators are taking. Carriers seem to think that, as long as they can adopt Wi-Fi and WiMAX earlier and develop a better business model than the independents, these will be technologies that they can turn to their advantages. For mobile operators, there is a doublededged sword. WiMAX is particularly disruptive because no physical last/first mile installation is required and the BS will cost roughly under $ using commodity standard hardware. As with Wi-Fi hot spots, fixed and mobile operators will have an equal interest in extending their networks through WiMAX, and ensuring that any revenues lost to 3G and wired services are at least preserved within the company. But WiMAX also gives the opportunity for small, alternative operators to enter the game. WiMAX deployment will follow a twostage development. Once mobility and broadband have been combined in step two in the form of integrated CPEs in 2006, WiMAX will coexist alongside Universal Mobile Telecommunications System (UMTS). This could represent what everybody is looking for: the killer application or the deadly (winner) combination. UMTS Node B Conclusions UTRAN IP/MPLS/Ethernet UMTS RNC Within five years, WiMAX is expected to be the dominant technology for wireless networking. By that time it will be fully mobile as well as providing low-cost fixed broadband access that will open up regions where access has so far not been practical. As the cellular operators move to IP-based 4G systems, they will embrace WiMAX as they are doing with the far more limited Wi-Fi. WiMAX will be the catalyst for the smart operators, with some of the small independents failing to the large players, still hunting for a more profitable revenue stream than 3G. However, there are few better examples of technologies that have promised so much, yet delivered so little over the years as long-range fixed wireless (for example, Bluetooth, FSO). At the same time there is a plethora of wireless standards emerging in 2004 from the IEEE and IETF, but only WiMAX seems to addresses all the key elements that are needed to make high-end wireless a reality, and which existing proprietary last/first mile and BWA technologies have failed fully to provide: a single standard for fixed broadband access and mobility, wireless WLAN backhaul, low cost of deployment, high scalability and the support of vendors with the power to drive the standard forward rapidly. Will WiMAX be dominated by non start-ups such as Intel and Nokia like it happened with Wi-Fi? Specialisation and the ability to integrate the product into a turnkey solution will be the key to success. In the following months there are expected to be consistent partnerships between infrastructure suppliers. Some of them have already been announced in press releases. Vendors of fixed CPEs will need to adapt to a market where the line between fixed and mobility applications is increasingly disappearing. Now, the question for the 44th FITCE Congress is: With the first WiMAX certified products due by the beginning of 2005, will the technology deliver on its promise? 1 Marks, Roger, B. (Chair, IEEE Working Group on BWA). IEEE Standard for Global Broadband Wireless Access. ITU Telecom World 2003, Geneva, Oct WiMAX Forum web site Gabriel, Caroline. WiMAX: The Critical Wireless Standard. BluePrint Wi-Fi Monthly Research Report, Oct Maravedis Inc. Broadband Wireless Intelligence web site. Acknowledgement Tables 1, 2, 3 and 4 courtesy of WiMAX Forum and Intel. Copyright Biography Cristian Patachia Sultanoiu Politehnica University, Bucharest Cristian Patachia Sultanoiu received his B.S degree in Electronics and Telecommunications (Computer Science and Information Engineering) from the Politehnica University, Bucharest, Romania, in He received his M.S. degree in Optical Fibre, Fixed and Mobile Radio Communications from Politehnica University, Bucharest in He is now a Ph.D. student in the same department. While working toward his Ph.D., he has also been working for Orange Romania as Transmission Expert in the Data Network Technology department. He is currently managing several projects examining future markets for voice and data VPN services and developing strategies and architectures based on BWA networks to meet future customer needs. His research interests include QoS in wireless networks including the new IEEE standard, NLOS wireless access networks and MAC enhancements for IEEE future networks. He is an AITR member and FITCE candidate member. 116