Emerging Technologies in Mobile and Wireless Data Network Evolution Proceedings of the Research Seminar on Telecommunications Business

Helsinki University of Technology Publications in Telecommunications Software and Multimedia Teknillisen korkeakoulun tietoliikenneohjelmistojen ja multimedian julkaisuja Espoo 2003 TML-C11

Editor Sakari Luukkainen

ISBN 951-22-6622-9 ISSN 1455-9749 2 Helsinki University of Technology Telecommunications Software and Multimedia Laboratory P.O. Box 5400 FIN-02015 HUT Tel. +358-9-451 2870 Fax. +358-9-451 5253

Helsinki University of Technology Telecommunications Software and Multimedia Laboratory Publications in Telecommunications Software and Multimedia Teknillisen korkeakoulun tietoliikenneohjelmistojen ja multimedian julkaisuja TML-C11 Espoo, 2003

Emerging Technologies in Mobile and Wireless Data Network Evolution Proceedings of the Research Seminar on Telecommunications Business II, spring 2003

Editor: Sakari Luukkainen (Lic.Tech.)

Keywords: mobile networks, wireless neworks, telecommunication business & investments, UWB, AdHoc, UMTS, WLAN, SDR, CDPD, EDGE, MMW, 60 GHz, MPEG-4, MVAS, Wireless Instant Messaging, Mobile Entertainment

The articles have been written by the students of the course T-109.551 Research Seminar on Telecommunications Business II in the spring 2003. The authors have full copyright to their articles. Technical editing by Eino Kivisaari. http://www.tml.hut.fi/Studies/T-109.551/2003/Proceedings.pdf ISBN: 951-22-6622-9 ISSN: 1455-9749 URN:NBN:fi-fe20031401

3 Preface

Mobile and wireless technologies are now in a discontinuous phase. As cellular mobile networks evolve from enhanced GSM to next generation systems by providing faster data connections, personal area networks (PAN) and wireless local area networks (WLAN) are also gaining popularity. Several wireless technologies are emerging, the success of which is unclear due to complex market dynamics, technology development and risks related to both. WLANs have already proved their effectiveness in the indoor business usage by extending the evolution of fixed LANs but the transition from voice to data in mobile networks has so far been slow despite the high expectations. Do we see in the outdoor data usage a complementary development of wireless and mobile technologies or does a single radio interface take a dominant role in the future like in the case of GSM? The experiences gained so far indicate that it is very difficult to predict the demand of new technologies and related services. Nevertheless, the launch of a service or R&D project usually requires large investments which imply big risks. The development of end user behaviour is more incremental than discontinuous. That is why it seems that we need parallel incremental evolution of services during the revolu- tion of radio related technologies. Another important issue is to realize that each emerging radio interface has its own strengths and weaknesses both in economic and technical terms. It is likely that the future network environment is a heterogenic multi-radio network which provides the end user a single subscription with differ- ing service quality depending on the geographic area in question. This publication is a collection of research reports written during the course Re- search Seminar on Telecommunications Business at Helsinki University of Tech- nology. The course is especially designed for students taking Telecommunications Management for their major but is as well suitable for all students that like to develop their techno-economic analysing skills in the telecommunications area. The aim of the Telecommunications Management major is to help the students to understand the structure and dynamics of economic life and industry with a special focus on the telecommunications by combining business and technology studies. The course provides an opportunity to rehearse scientific writing and presentation skills. The goal of the spring 2003 seminar was to investigate the challenges service providers face when updating their network infrastructure to offer emerging mobile and wireless data services taking into consideration both business and fast developing technology requirements and possibilities. One major challenge is to move from a technology-oriented to a business/service-oriented approach. As a result of hard work we have now a collection of interesting papers of several possible technological scenarios in the mobile and wireless data network evolution. I want to thank all the contributors from excellent papers as well as lively discus- sions during the seminar sessions.

Sakari Luukkainen Espoo, June 5th.2003

4 Contents

Sami Vesala and Katja Koivu Positioning EDGE in the Mobile Network Evolution ...... 6

Guoyou He UMTS and Its Market Analysis ...... 16

Yaojin Yang UMTS Investment Study ...... 25

Feng Yue CDPD Service ...... 32

Tapio Sokura 5 GHz WLAN ...... 37

Riku Honkanen WLAN Operator Cases ...... 42

Henrik Petander and Olli Savolainen Ad Hoc Networking as an Internet Access Technology...... 51

Olli-Pekka Isola and Ari Haapaniemi Market Strategy Evaluation of Ultra Wide Band Technology ...... 57

Eino Kivisaari 60 GHz MMW Applications...... 61

Kai Kuikkaniemi Introduction to Software Defined Radio ...... 66

Mihai Burlacu and Sonja Kangas MPEG-4 Technology and Business Strategies...... 71

Kimmo Palletvuori Changing MVAS Environment ...... 75

Qinghai Li MMS and WIM Technology and Business ...... 81

Min Chen Mobile Entertainment ...... 89

5 Positioning EDGE in the Mobile Network Evolution

KATJA KOIVU SAMI VESALA TELECOMMUNICATIONS SOFTWARE COMMUNICATIONS LABORATORY AND MULTIMEDIA LABORATORY HELSINKI UNIVERSITY OF TECHNOLOGY HELSINKI UNIVERSITY OF TECHNOLOGY P.O. BOX 3000, 02015 HUT, FINLAND P.O. BOX 5400, 02015 HUT, FINLAND [email protected] / [email protected] [email protected] / [email protected]

ABSTRACT 2 CELLULAR NETWORK EVOLUTION In this paper, we present EDGE technology along with its Cellular radio networks are generally divided into three business opportunities. EDGE gives GSM the capacity to generations. handle services for the third generation of mobile net- Analogue cellular systems, such as Nordic Mobile Tele- works. EDGE was developed to enable the wireless trans- phone (NMT), are considered to be the first generation of mission of large amounts of data at a higher speed than cellular technologies. before. EDGE will allow GSM operators to use existing GSM radio bands to offer IP-based multimedia services The second generation is the present digital network gen- and applications. Implementing EDGE will be relatively eration which includes systems like Global System for easy and will require relatively small changes to network Mobile communications (GSM), Digital Cellular System hardware and software. (DCS), Digital Advanced Mobile Phone System (D- AMPS), and Interim Standard –95 (IS-95). The second generation includes also enhancements to GSM: High 1 INTRODUCTION Speed Circuit Switched Data (HSCSD), General Packet Radio Service (GPRS) and Enhanced Data rates for GSM EDGE technology gives GSM the capacity to handle ser- Evolution (EDGE). These enhancements are called the vices for the third generation of mobile networks. EDGE generation 2G+ or 2,5. was developed to enable the wireless transmission of large amounts of data at a higher speed than before. According to International Telecommunications Union (ITU) specifications, the third generation cellular networks EDGE will allow GSM operators to use existing GSM will offer data transmission speeds up to 2Mbps. Univer- radio bands to offer IP-based multimedia services and ap- sal Mobile Telecommunications System (UMTS) is one plications at theoretical maximum speeds of 384 kbps with of the mobile communications systems being developed a bit-rate of 48 kbps per timeslot and up to 69.2 kbps per within the ITU framework known as International Mo- timeslot in good radio conditions. bile Telecommunications IMT-2000. Implementing EDGE will be relatively easy and will require relatively small changes to network hardware and software as it uses the same TDMA (Time Division Mul- GSM 9.6kbps UMTS tiple Access) frame structure, logic channel and 200 kHz 2Mbps carrier bandwidth as today’s GSM networks, which allows existing cell plans to remain intact. HSCSD 57.6kbps

This paper focuses on firstly studying the technology be- GPRS hind EDGE and secondly on studying business related 115kbps EDGE issues. 384kbps

This paper is organized as follows. Sections 2 and 3 1999 2000 2001 2002 2003 present cellular network evolution along with mobile service evolution towards 3G. In Section 4, EDGE technol- Figure 2.1. Evolution paths of GSM towards third ogy is studied. Section 5 presents vendors EDGE strate- generation networks gies. In Section 6, EDGE enabled services are presented. Section 7 shows terminal availability. In Sections 8 and 9 The different GSM evolution paths are shown in Figure EDGE investment costs and revenues and EDGE invest- 2.1. The data rates are the maximum data rates theoreti- ments strategies are studied. Finally in Section 10, the cally provided by different systems. In reality, maximum future role of EDGE is discussed. Concluding remarks data rates are achieved only in very limited circumstances, are provided in Section 11. if at all.

6 3 EVOLUTION OF MOBILE SERVICES ture messaging part of their lifestyle is service pricing. MMS-compatible phones are expensive and users will Voice services are still the most important services pro- only be persuaded to buy these camera phones if they can vided by mobile communications networks. However, the afford to use them. The widespread usage of SMS text earlier presented enhancements to ordinary GSM tech- messaging has been enabled by service pricing which is nology bring new possibilities for various data services. both easy to understand and fairly cheap. [1][2] At the same time, the importance of other than voice services grows rapidly. The development of mobile data services follows the evo- 3.2 HSCSD and GPRS enabled services and data lution path of cellular technologies. The first generation rates in practice analogue cellular systems offered extremely slow and un- Today, HSCSD and GPRS connections are mainly used reliable data connections and identification methods were for accessing email, getting information from the Internet, not well developed. web surfing in general, entertainment (e.g. downloading The second-generation digital cellular systems made an video clips, music, etc.), banking and shopping. The data improvement to the data services and data rates. In addi- rates achieved by using HSCSD and GPRS with termi- tion, Subscriber Identification Module (SIM) cards in nals available on the market are quite similar. GSM phones improved security and enabled, for example, The current terminals and networks do not support much safe bank connections and using of cellular phones for over 40 kbps data rates in practice – this means slower money transactions. and more unstable connections than ordinary fixed-line The Short Message Service (SMS) provides guaranteed modems can offer. delivery of small data packets even if the phone is switched HSCSD connections are more stable compared with GPRS off when the message is first sent. connections because of the circuit-switched nature of Now, second-generation services offer higher bit rates and HSCSD. Even though, it is not assured, that a HSCSD packet-switched connections. The development path ad- user gets all the three timeslots his terminal can carry for vances towards UMTS and third generation services that the downlink traffic. In many cases, circuit-switched offer the ground for many high-speed services. In addi- speech is prioritized over HSCSD and GPRS traffic and tion, wearable computers and totally computerized homes thus timeslots first allocated for data are allocated for can be a part of everyday life after a few years. Wireless speech on the go. In the current GPRS networks only CS- Local Area Network (WLAN) products and other pos- 1 and CS-2 coding schemes are used. The average C/I sible wireless network applications can have a remark- from tested networks leads to 11,5 kbps/timeslot. There- able role in parallel with advanced cellular network ser- fore in practice, the networks offer “best effort” service vices. quality and slow connections. Giving data users enough capacity is also a pricing issue. Adding capacity to the network is always an extra invest- 3.1 Multimedia Message Service ment. Operators in Finland have different pricing strate- The multimedia message service (MMS) has become a gies and thus also their network parameters for GPRS traf- significant issue for mobile operators future growth strat- fic differ from each other. egies. MMS is expected to be the most important service for operators, content providers and service providers since MMS will provide them with a new source of rev- 4 EDGE TECHNICAL FUNDAMENTALS enue now and in the 3G markets. Enhanced Data rates for GSM Evolution (EDGE) is a ma- The key to MMS is to maintain the fundamental features jor enhancement to GSM/GPRS data rates and it improves that have made SMS a success story, while offering con- the GSM air-interface performance significantly. EDGE sumers a more versatile and personal experience. MMS offers improved data rate through optimized modulation will enable consumers to send and receive multimedia (8-PSK) and it introduces a large number of channel cod- messages between mobile terminals as well as between ing schemes along with Incremental Redundancy (IR), terminals and content servers. MMS messages combine Link Adaptation (LA) enhancements and in the near fu- image, sound and text, and even animation and video. The ture adaptive multirate (AMR). camera phones, which are currently available, do not pro- duce pictures, which are bigger than 100Kbytes. The new modulation and the possibility to adapt the transmission rate to channel quality are the core of the EDGE Now, mobile operators have started to push MMS ser- concept. Introducing EDGE in a GSM network does not vices seriously. At the start of 2002, only a single opera- imply changes in the basic architecture. In any case, modi- tor, Norway’s Telenor, had launched MMS-based picture fications of the Mobile Station (MS), Base Station (BTS) messaging. By November 2002, over 60 operators world- and Base Station Controller (BSC) are needed, which wide were offering picture messaging. means, among other things, software and hardware up- One of the key factors that will determine the answer to grades in circuit- and packet-switched parts of the net- whether or not large numbers of consumers will take pic- work.

7 EDGE offers both circuit- and packet-switched connec- (MCS) are introduced. Classes MCS-1 – MCS-4 use the tions depending on the platform it is implemented in. The basic GSM 0.3 GMSK modulation, whereas classes MCS- scope of the EDGE phase 1 standard is to increase GPRS 5 – MCS-9 use the new 8-PSK modulation. Table 4.1. bit rate, improve GPRS link quality control (EGPRS) and shows EGPRS modulation and coding schemes along with to offer high circuit-switched data rate with fewer timeslots their maximum throughputs. and fast power control (ESCD). The scope of the EDGE phase 2 includes supporting real-time services over EGPRS. GSM networks have already offered advanced data services from single SMS and circuit-switched 9,6 kbps data services to 64kbps HSCSD and 160 kbps (theoretical speed) GPRS. With EDGE, the data rate offered by the original HSCSD or GPRS networks can triple. [3]

4.1 8-PSK modulation in GSM/EDGE standard EDGE is specified to reuse the channel structure, channel width, channel coding and the existing mechanisms and Table 4.1. EGPRS modulation and coding schemes functionality of GSM, HSCSD and GPRS. The enhancement behind tripling the data rates is the introduction of the new modulation type. Another improvement that has been made to EGPRS stan- The modulation type that is used in GSM is the Gaussian dard is the ability to retransmit a packet that has not been minimum shift keying (GMSK), which is a kind of phase decoded properly with a more robust coding scheme, modulation. EDGE introduces the octagonal phase shift whereas for GPRS re-segmentation is not possible. In keying (8-PSK) modulation in addition to the existing GPRS once packets have been sent, they must be retrans- GMSK, see Figure 4.1. mitted using the original coding scheme even if the radio environment has changed.

(d(3k),d(3k+1),d(3k+2))= (0,1,0) (0,0,0) 4.3 Link adaptation (0,1,1) EGPRS uses automatic link adaptation (LA). LA is used (0,0,1) (1,1,1) to select the best MCS for the radio link conditions. LA uses the radio link quality measured either by the mobile (1,0,1) (1,1,0) station or by the base station to select the most appropriate modulation and coding scheme for transmission of (1,0,0) packets. Each modulation and channel coding class is optimized for a certain range of C/I (interference) values, Figure 4.1. 8-PSK signal constellation principle outside of which the data rate no longer increases together with the C/I value, but saturates. LA algorithms compare The number of symbols sent within a certain period of the estimated channel quality to threshold values and that time, the symbol rate, remains the same as for GMSK but leads to optimized throughput. In reality, the link adapta- an 8-PSK signal is able to carry three bits instead of one. tion may not be close to the ideal situation where the maxi- The total data rate is therefore increased threefold. mum data rate is (as a function of the C/I curve) achieved An 8-PSK modulated signal is more sensitive to errors by switching channel coding class “on the go”. [3][4]] and thus the highest data rates can only be achieved with limited coverage. GMSK modulation is more efficient 4.4 Incremental redundancy under very poor radio conditions and therefore EDGE coding schemes are a mixture of both GMSK and 8-PSK. Another way to choose the optimal channel coding class [3][4] is to use the incremental redundancy technique (IR). In- cremental redundancy initially uses a coding scheme with very little error protection (such as MCS-9) and without 4.2 Enhanced general packet radio service (EGPRS) consideration for the actual radio link quality. When in- Enhanced general packet radio service (EGPRS) is build formation is received incorrectly, additional coding is on top of GPRS. transmitted and the resent information is soft combined in the receiver with the previously received information. Four different coding schemes are defined for GPRS (CS- IR adjusts the code rate of the transmission to true chan- 1 to CS-4). Each has different amounts of error-correct- nel conditions with incremental transmissions of the re- ing coding that is optimized for different radio environ- dundant information until the decoding is successful. For ments. For EGPRS nine modulation and coding schemes the mobile stations, incremental redundancy support is

8 mandatory in the standard. The information about the ra- user, while allowing operators to efficiently manage the dio link is not necessarily to support incremental redun- available spectrum. The existing GERAN radio protocols dancy. IR gives additional 2-3dB to the radio link. [3][4] need to undergo significant modifications, and this will increase the complexity of radio interface protocols. In addition, standardization of the GERAN Rel5 should sup- 4.5 Enhanced circuit switched data (ECSD) port a true multi-vendor environment and GSM/EDGE Enhanced circuit switched data (ECSD) is based on the radio access should be backwards compatible, i.e. sup- current HSCSD is GSM networks. The ECSD architec- port of services for pre-Rel5 terminals must be ensured. ture is mainly based on HSCSD transmission and signal- [3][4] ing. ECSD does not increase the maximum 64 kbps data rate 4.7 GERAN Rel5 features of HSCSD but it makes the network more efficient: the In the 3rd Generation Partnership Project (3GPP) Rel5 same data rates can be achieved with allocation of fewer overall, the most significant new functionality is the timeslots and simpler MS implementation. Internet multimedia subsystem (IMS). From the GERAN Circuit-switched data connections up to 64kbps are suffi- perspective, the support for the IMS services implies in- cient for providing various transparent and non-transpar- troduction of the Iu interface, and definition of the header ent services, e.g. interworking with audio modems and adaptation mechanism for the real-time protocol (RTP), ISDN at various data rates and various video based ser- user datagram protocol (UDP), and Internet protocol (IP) vices ranging from still image transfer to traffic. Rel5 includes also major enhancements for speech: videoconferencing services. wideband AMR speech for enhanced speech quality, half- rate 8-PSK speech for improved speech capacity, and fast In the future, Enhanced Adaptive Multi Rate codec power control for speech. In addition to the (EAMR) enables the transfer of high-quality speech and abovementioned enhancements, Rel5 implies location music. The same restrictions apply to EAMR connections service enhancements for Gb and Iu interfaces and inter- as apply to ECSD, as EAMR is also circuit-switched. BSC and BSC/RNC network assisted cell change (NACC). [3][5] [3]

4.6 EDGE evolution towards GERAN Rel5 4.8 GERAN Rel5 system architecture GSM/EDGE radio access network (GERAN) Rel5 in- To connect to the WCDMA/GPRS core networks, cludes a definition of enhancements to the GPRS radio GERAN will use the Iu interface, as shown in Figure 4.3. link interface and will provide support for conversational The Iu interface connects to the circuit-switched domain and streaming service classes as defined for WCDMA. (Iu-cs) and to the packet-switched domain of the core net- With the adoption of the UMTS Iu interface and the work (Iu-ps). GERAN also connects to the second-gen- UMTS quality of service (QoS) architecture in Rel5, eration core network nodes by using the A and Gb inter- GERAN and UTRAN can be efficiently integrated under faces. These interfaces remain intact in Rel5 to support a single UMTS multi-radio network. In addition, GERAN Rel’99 terminals. Iu-ps interface is not used for Rel´99 will include performance enhancements for existing ser- terminals because the functional split between the radio vices. access network and the core network differ substantially between Iu and A/Gb. GPRS/EDGE Radio Network

EDGE BS Network Subsystem

BTS BSC GERAN MSC/VLR PSTN BSS Iur-g HLR MS A UTRAN BTS BTS RNC Gb Um BSC GPRS-backbone MS BTS Iu BTS RNC SGSN GGSN IP Network

Iur-g UMTS Radio Network Core Network UTRAN RNC GSM/WCDMA Figure4.2. A simplified model of the combined GSM Core Network GPRS/EDGE and UMTS network [6]

In general, the goals and impacts of GERAN Rel5 speci- Figure 4.3. GERAN architecture in Rel5 fication are to enable GERAN to the same 3G CN (core The radio interface Um between GERAN and the mobile network) as UTRAN creating first steps towards efficient station is based on the Rel99 radio interface link specifi- resource optimizations in multi-radio networks, and to en- cations. However, several modifications are needed on able GERAN to provide the same set of services as radio link protocol layers in order to provide adequate UTRAN, making the radio technology invisible to the end- radio bearers for real-time services. These modifications

9 imply to support for cell reselection for packet-switched domain, separation of user and control planes, and trans- TRX parent modes in the radio link protocol layers. [3][4] requirements Interference matrix Coverage Propagation • co-channel Frequency Analysis estimations • adjacent plan 4.9 Modifications to the GSM network imposed by channel Separation EDGE constraints The implementation of GSM EDGE requires basically only TRX change to EDGE TRXs in the GSM base sta- Figure 4.5. Frequency planning principle tions and software updates to GSM BSC and GPRS IP- backbone. A bigger investment would most probably be the upgrade of Abis interface from 16 kbit/timeslot con- EDGE deployment doesn’t bring dramatic changes to ra- nection to EDGE capable 64 kbit/timeslot connection. [6] dio network planning with GPRS. Main concerns are the allocation of capacity and steering of traffic to wanted The impact of EGPRS on the existing GSM/GPRS net- layer/cell/TRX. Changes to transmission capacity will be work is limited to the base station system due to the mi- needed, if larger scale EDGE deployment per cell/area is nor differences between GPRS and EGPRS. A new trans- done. ceiver unit capable of handling EDGE modulation as well as new software that enables the new protocol for packets The easiest way to implement EDGE from the network over the radio interface in both the base station and base planning point of view is the TRX replacing strategy, station controller. The core network remains intact. [4] where new frequency plan is not mandatory. The replacing can be done for every 1-3rd site to achieve coverage and EDGE services e.g. hotspots or rural area can be selected for EDGE, but with limited amount of data through- OSS put.

2G SGSN GGSN Higher data amounts with EDGE can be offered if it is BTS G implemented by bringing an additional EDGE TRX dedi- n BSC IuIu cated to data usage to (some of) the cells in the network A-bis A MSC and/or by reserving more timeslots for the use of EDGE data users from the TRXs. However, that leads to decrease BTS in the GoS experienced by the speech users. In real life

GSM/EDGE these actions are not always possible to perform and they will require significantly more implementation and planning work. Figure 4.4. EDGE implementation [7] In order to utilise EDGE performance in full, a totally new frequency plan and possibly new GSM cell structure In Ericsson case, EDGE is compatible with recent equip- are required. ment. If an operator has an Ericsson RBS 2000 macro base station from 1995 or later, it is easy to take on EDGE. Some additional hardware using plug-in transceivers, and 5 VENDORS’ EDGE STRATEGIES new software that can be installed remotely is all that is Due to the delays in UMTS implementation compared to needed for operators to start offering high-quality Mo- the early predictions, most vendors have taken EDGE back bile Internet services over their existing infrastructure. [8] to the table. EDGE as a technology was firstly developed In Figure 4.4, the elements for EDGE implementation are to support the GSM evolution towards 3G in especially shown. US markets. As UMTS hype began in 1999/2000 EDGE was put to less priority among most of the vendors, be- 4.10 EDGE radio network planning compared with cause UMTS implementation seemed to happen so fast GSM/GPRS planning and with large scale that it was natural to shift all focus to support this. If we think the implementation of EDGE of the radio network planning perspective, the same principles as in the After the enormous UMTS license fees most of the op- GSM/GPRS network planning apply. As in GPRS, EDGE erators’ capabilities to invest in to the networks were de- performance is dependent on the achievable C/I (and creased significantly. At the same time making the UMTS RXlev) in the network. The most effective means to gain technology work needed more work from the vendors than high performance in good radio conditions is to come up anticipated, which together with the operators decreased with a optimized frequency plan. Frequency plan optimi- investment capabilities caused the delays in UMTS imple- zation can make a significant difference for the achiev- mentation. EDGE was again an issue, because the needed able throughput. investments on it are a fraction of those needed for UMTS and the end user performance is quite close to UMTS in the beginning. Of course the capacity offered by UMTS

10 is enormous compared to EDGE, but as there is no pres- 6 SERVICES ENABLED BY EDGE sure on the capacity side for the operators (because the The Release 99 EDGE implementation does not offer sig- data traffic haven’t proved to increase still) UMTS ca- nificant new possibilities for services compared with the pacity is not really needed yet. current HSCSD and GPRS networks. When the US operators, for example AT&T, started to As mentioned before, circuit-switched data connections implement EDGE capable HW as they decided to go for up to 64kbps are sufficient for providing various trans- the GSM evolution towards 3G rather than IS-95 based, parent and non-transparent services, e.g. interworking with there was suddenly a need for the operators to start mak- audio modems and ISDN at various data rates and vari- ing EDGE terminals as well. As the biggest reason for the ous video based services ranging from still image trans- US operators to choose GSM based system is the roam- fer to videoconferencing services. Packet-switched con- ing traffic gained globally, there is a clear need to have nections are optimal for bursty data traffic, e.g. web brows- EDGE happening in the Europe and Asia as well. This ing and email. was the reason why EDGE marketing started again with full steam for European operators as well by the biggest In Rel5 UMTS 3GPP traffic classes are enabled in EDGE vendors. and thus 3G services delivery across all frequency bands and bearers becomes possible. Handovers across GSM/ The biggest two vendors, Nokia and Ericsson, are the most EDGE/WCDMA are enabled from the start. However, active with EDGE marketing. Nokia has been clearly the there are still uncertainties in standardization of Rel5 and market maker in Europe and Asia Pacific for EDGE. It’s the Iu interface. of course in interests of all the vendors to make EDGE a success, but due to its strong position in terminal market, When compared to GPRS phase 1 QoS classification, the Nokia is in better position to drive the market than it’s QoS grouping of UMTS release 99 takes into account the competitors. Ericsson has clearly taken a follower role in applications that will become available through the in- EDGE market, focusing clearly on driving the UMTS creased data rates of UMTS and EDGE. The main distin- market. This is also partly due to the difficult financial guishing factor between the traffic classes is the sensi- situation of Ericsson currently, where the ability to take tiveness of applications, as presented in Table 6.1. [6] risks in new market areas is limited. When the two vendors are compared from the point of view of needed in- Traffic class Example of application Fundamental characteristics crements to the legacy network infrastructure (to make Conversational Voice and video Preserve time relation between EDGE possible), Ericsson is in stronger position than class telephony information elements, low delay Nokia with better applicability of the older infrastructure Streaming Real time Preserve time relation between, in the field. class streaming video low level retransmission Interactive Web browsing and real Preservation of content, All the other noticeable GSM network vendors (Siemens, class time control channels retransmission, "request response" Background Downloading of files Delay insensitive, preservation of Alcatel, Motorola, Nortel) have taken a reactive role with class and email content, retransmission EDGE, waiting for the market to start up. Outside USA there has been little marketing done for EDGE by these Table 6.1. QoS classes for UMTS and EDGE vendors and they are clearly waiting and seeing whether the big vendors (mainly Nokia) can have the market cre- Conversational and streaming classes are mainly intended ated for EDGE and then jumping on board. Of course to be used to carry real-time traffic flows. The main dif- they all have EDGE infrastructure and terminals as well ference between them is the delay sensitiveness of the in their road-maps, but they are not put into number one classes. Interactive class and background class are mainly priority and committed on. meant to be used by the Internet type applications e.g. The problem with EDGE is for a network vendor that it web browsing and e-mail. Due to looser delay require- has been earlier positioned as 3G technology thus com- ments compared to conversational and streaming classes, peting with the UMTS market. So, as the potential UMTS both provide better error recovery by means of retrans- market is clearly bigger than EDGE, it has been decided mission. by most vendors not to drive EDGE strongly towards their

Voice customers. This could have an negative effect on the Voice Infotainment UMTS sales. The answer is to position EDGE more as a Video telephony Communication

Conversational FAX Transaction WCDMA Collaborative working services enhancement to existing GPRS networks and to co-exist Video streaming Advertising Corporate with UMTS. solutions

Streaming Audio streaming Conversational Gaming Currently as first Nokia and then Sony-Ericsson have com- Corporate Data Access E-mail EGPRSWEB Browsing mitted to bring EDGE terminals to Europe-Asia GSM Interactive Conversational WAP Applications bands as well, it seems that the market is clearly starting. MultimediaGPRS Messaging Audio clip downl. The availability of terminals and thus necessary penetra- Short Background Messaging Conversational Video clip downl. Data rate tion is in vital role in possible EDGE success. [9][10][11] 0 8 16 48 128 473 2048

Figure 6.1.Service QoS Requirements for Bearers, Data rates and services 2003 [7]

11 The main difference between interactive and background The size of the investments on EDGE depends on the op- class is that interactive class traffic will have higher pri- erator. The worst case is that large part of the network ority in scheduling than the background class traffic. This infrastructure is old enough to not support easy EDGE means that background applications may use transmis- implementation. The capability of the infrastructure de- sion resources only when other applications do not need pends of the network vendor. In some cases the base sta- them. [3] tions must be fully replaced by newer ones before EDGE can be implemented. In typical cases only EDGE capable Although the conversational class is specified in the QoS TRXs and BSS software must be implemented along with classes of UMTS release 99, the most delay-critical ap- changes to Abis interface capacity. If the network infra- plications such as speech and video telephony will be structure is new enough, the base stations can be already carried on circuit-switched bearers in the first phase of equipped with EDGE capable TRXs. Then only software the third generation mobile networks. Later it will be pos- and enhanced transmission capacity must be implemented sible to support delay-critical services as packet data with and thus costs can be kept quite minimal. QoS functions. Different QoS requirements of services are shown in Figure 6.1. In typical cases of operators’ network evolution the changes are made concurrently as much as possible. This means that for example if the network infrastructure is 7 TERMINAL AVAILABILITY old enough to require new base stations if EDGE is considered, can the changes planned so that UMTS/EDGE The first terminals for EDGE will be on the market by capable base stations are introduced at the same time. This 2H2003 by Nokia and Motorola. These are aimed for the lowers the needed investments (and the operational ex- US market, but will have also European GSM frequen- penses) for EDGE, which would separately be quite enor- cies available. For example the Nokia 6200 will have GSM mous. Similarly EDGE capable infrastructure can be 1900/1800/800 frequencies and support for MCS1-9. With moved to replace older infrastructure in the needed areas the same timetable Nokia will also introduce an EDGE if available. This also makes the investments lower. De- terminal with GSM 900 frequency available. pending on the situation of the operator, the following The Motorola t725 will have also both European frequen- costs are related to EDGE implementation: cies imbedded and supports MCS1-9. The t725 will be available by 2H2003. - EDGE capable GSM/GPRS base stations Sony-Ericsson has also said that it will bring EDGE capable terminals for both European and US markets in the - EDGE capable TRXs second half of 2003. An interesting product will be also - EDGE capable BSS software the PC-card with GPRS/EDGE from Sony-Ericsson, which will be available also in the second half of 2003. - Possible capacity upgrades to BSC The most important fact when EDGE terminals are con- - Enhancements to Abis capacity sidered is the information given by Nokia, that it will in- - NMS/OMC changes clude EDGE to all of its GPRS terminals that are introduced after 6/2003. Nokia has also said that it will have - Possible upgrades to GPRS core network EDGE included in all of its terminal categories from the - Network planning costs (site configuration planning, beginning of 2004. This is extremely good news for frequency planning etc.) EDGE, since it almost positively ensures that the terminal penetration will start to develop and that other ven- - Operational costs of implementation dors will join in manufacturing EDGE terminals. It also gives operators a positive signal to include EDGE in their Since the investments needed for EDGE are highly de- network evolution strategy as a realistic option. [9][11] pendant on operators network, strategy and cost structure and network vendor’s capability and pricing towards a specific operator, it’s quite impossible to give generic cases 8 INVESTMENT COSTS AND REVENUES of the needed total investments. It can however be said EDGE is a further development of the GSM/GPRS net- that they are a fraction of that needed for UMTS. EDGE works and thus it can be integrated in the already estab- can be implemented to every third site for example, so it lished GSM/GPRS networks with relatively low invest- enables lots of different capacity/coverage strategies, ment. Beside from the hardware and software upgrades, which can be used to optimize the costs involved. Some it only affects the network by increasing capacity and data practical cases have shown that the pricing for EDGE rates. The cost of operation will not increase. Operators TRXs and BSS software is quite similar or only a bit higher can deploy EDGE using the existing GSM spectrum. to that of GPRS equipment. This can also be due to the EDGE has higher spectral efficiency than GSM/GPRS fact that the vendors need the reference networks up and and thus there is free capacity for carrying more data and running, which usually means that the margins for the sales voice traffic and for serving more subscribers. are kept lower than usual.

12 As the EDGE capable terminals reach a feasible penetra- EDGE as a stepping stone can require EDGE to be imple- tion percentage, the data traffic is more economical to be mented over the network, which of course will make the served with EDGE rather than GPRS, this is because the investments bigger as well. If the network is built directly capacity provided by EDGE is almost 3-fold compared to support EDGE then the investments can be made to GPRS, with relatively small investments needed. This smaller. Of course EDGE can be implemented only to fact also allows more capacity for speech service, taken e.g. cities and GPRS elsewhere, which lowers the incre- that the network is parameterized accordingly. This en- mental costs. ables greater revenues for an operator. Also the higher throughput offered by EDGE for the users can be priced higher than the conventional GPRS. Later if the GERAN 9.2 GSM operators with 3G license offers same QoS functions as UTMS it will create even EDGE and WCDMA can also be complementary 3G tech- more possibilities to generate revenue from the users. [11] nologies that together will sustain the operator’s need for nation-wide mobile data and speech capacity during the expected traffic boom. They are both IMT-2000 capable 9 EDGE INVESTMENTS STRATEGIES radio access technologies in different frequency bands. They can both provide 3G services for the end-user, ac- Different second-generation systems have different evo- cessing a common core network, given that the Iu-inter- lution paths towards third generation IMT-2000 services. face is standardized to be supported by the GERAN as Network operators that are not granted UMTS licenses well. can implement EDGE to offer IMT-2000 alike services. However, an operator with UMTS license may still de- Possible business cases for the operator with a UMTS ploy EDGE to create a wireless data market before third license are for example the following: generation CDMA systems are launched or use EDGE in EDGE used as a complementary solution, to different areas where there is no UMTS service. coverage areas. In this case EDGE is implemented e.g. to more rural areas and UMTS to the cities and sub-urbans 9.1 GSM operators without 3G licenses and main roads. This case makes it possible to offer high- data rate services also in the areas where UMTS cover- For those operators without 3G license, EDGE offers a age is not available. The investments of EDGE imple- pretty straightforward business case as a stepping stone mentation can be made lower by guaranteeing that the before UMTS. Of course an operator can choose not to infrastructure in the “EDGE areas” is the most EDGE ca- go for UMTS at all and offer high data rate services with pable one, if possible by swapping. GPRS/EDGE network. This type of stepping stone approach is currently used by some of the US operators (such This case requires though that there will be multi-mode as AT&T), which have not decided the UMTS bandwidth terminals available quite rapidly (i.e. UMTS/EDGE/ and are implementing GSM-EDGE networks as this is GPRS), because otherwise the experienced throughput can written. In Europe most of the countries have already be higher in the rural areas than in the cities, which is not granted their licenses so there are very few left to imple- very feasible. ment this strategy. In Asia, there is a technology standard EDGE and UMTS co-exist, for different user segments. war going on between CDMA2000, WCDMA and TD- In this case EDGE and UMTS are implemented to same SCDMA. In this area EDGE is clearly seen as a less at- coverage areas, but used to serve different user segments. tractive option. For example EDGE could be used to offer robust data for EDGE as a stepping stone to UMTS can be seen only corporate access and UMTS to offer fancy 3G services, feasible in markets with no strong UMTS commitment: such as video streaming etc. No licenses yet awarded or they are very inexpensive or This case will require quite large investments, because no present market push for 3G or if the GSM business is EDGE is to be implemented over the network or at least still developing. Besides US, this kind of situation exists to most locations. This case would also require that there e.g. in East-Europe. Although if the first UMTS launches would not be multi-mode terminals widely available very (e.g. Hutchison UK, Italy) are successes and UMTS ter- rapidly, so the segmentation could be made more easily minals come faster to market and are more strongly sub- through terminals. [11] sided than the EDGE terminals, then the window for EDGE feasible will become smaller rapidly for this type of strategy. The only clear situation is in the US, because 10 FUTURE ROLE OF EDGE they will not have UMTS licenses awarded for some time. In Figure 10.1, the current situation of EDGE globally is If this type of strategy is chosen by the operator, all the depicted. services can/will have to be planned on top of GSM- EDGE. This means that if for example streaming services are to be offered must EDGE standardization support this. In the first releases of EDGE the QoS will be similar to GPRS, which doesn’t enable similar services as can be offered by the UMTS.

13 make EDGE a niche technology. If UMTS success won’t happen in short term, due to for example technical difficulties, then EDGE has a better chance to become a widely deployed technology. In this case there will be wider range of EDGE terminals available before UMTS is deployed in larger scale. This will also bring in more EDGE operators as the investments on network technology are considerably smaller. The long-term success of EDGE is also dependant on the operators strategy on driving UMTS with terminal subsidization. This is easily the case in countries where the license terms are tight and demand a quite rapid and wide UMTS deployment. The most probable case is that EDGE will live alongside Figure 10.1. Global situation of EDGE in 2002. UMTS and will in most cases be used as a geographical complement to UMTS. The UMTS early success is not As can be seen from the figure, the only market areas in very likely, so it will give operators a chance to consider which public commitments to EDGE have been made is EDGE as well (which is already ready as a technology) the US and Asia Pacific. In Europe, which is the key mar- and invest still to their GSM networks. Also as the tech- ket area for a larger scale EDGE success, has not yet seen nology develops in the future, it will be more likely that any public EDGE commitments from the operators. Al- EDGE will be included in the terminals very cheaply in though the biggest vendors claim to have multiple con- the long run. This makes the multi-mode terminal manu- tracts for EDGE deliveries, no of them are public yet in facturing more feasible to the vendors. So it seems that Europe. EDGE will have a relatively good future ahead, especially The future role of EDGE depends very highly on the fol- in those countries, which don’t have too tight UMTS li- lowing: cense terms for the operators or have granted licenses cheap. In the long run the standardization must succeed to include the Iu-interface support to GERAN, otherwise - The success of EDGE in USA EDGE can easily be positioned as the “poor man’s UMTS” in the market as similar QoS and services cannot be of- - The availability of EDGE terminals fered. [9][10][11] - The early success of UMTS - Subsidisation of UMTS terminals 11 CONCLUSIONS - The completeness of EDGE standardization EDGE is relatively easy and cheap to bring to the operators network. It demands for significantly minor changes The biggest guarantee that EDGE will become one of the to the operational side as well. This means that it can be steps of GSM network evolution, which be implemented quite attractive for an operator, which have not made as well, is the availability of EDGE capable terminals. strong commitments to UMTS. This has already happened as Nokia & Motorola will bring EDGE uses 8-PSK modulation, which enables approxi- the terminals by the 2H of 2003. Basically this was obvi- mately 2,5 times the performance or capacity of GPRS. ous after the big US vendors started to implement GSM- This new modulation scheme requires new radio parts for EDGE networks. The only question mark is whether the terminals as well, which means that new terminals must EDGE will be widely deployed or will it become a niche be introduced from the vendors and the penetration of the technology. EDGE capable terminals grow enough before this capac- This depends much on the early success of UMTS in Eu- ity gain compared to GPRS can be fully utilized by the rope. If the first implementations in 1H2003 succeed, operators. At the moment there are commitments made which will bring in more network launches in 2H2003 from the biggest vendors to introduce EDGE terminals and 1H2004, then EDGE will more probably stay as a by 2H2003. niche technology deployed only to part of the networks. The size of the investments on EDGE depends on the op- This is because when UMTS succeeds, the focus of the erator. The worst case is that large part of the network industry is again shifted to UMTS and the investments on infrastructure is old enough to not support easy EDGE GSM networks will be minimized. Then EDGE will prob- implementation. The capability of the infrastructure de- ably experience similar situation as HSCSD did. Most pends of the network vendor. In some cases the base sta- probably there will be also operators, which will not go tions must be fully replaced by newer ones before EDGE for UMTS in the near future and for those EDGE offers a can be implemented. In typical cases only EDGE capable feasible solution. Altogether early UMTS success would TRXs and BSS software must be implemented along with

14 changes to Abis interface capacity. If the network infra- EDGE will be included in the terminals very cheaply in structure is new enough, the base stations can be already the long run. This makes the multi-mode terminal manu- equipped with EDGE capable TRXs. Then only software facturing more feasible to the vendors. So it seems that and enhanced transmission capacity must be implemented EDGE will have a relatively good future ahead, especially and thus costs can be kept quite minimal. in those countries, which don’t have too tight UMTS license terms for the operators or have granted licenses The strategy for an operator to choose on EDGE depends cheap. In the long run the standardization must succeed on many things. Most of it can be even personal (prefer- to include the Iu-interface support to GERAN. ences of decision making people) or contractual (towards vendors) reasons, which drive the network evolution. But in principle one deciding factor is the commitment towards UMTS and the license terms of the UMTS license. REFERENCES For those operators without 3G license, EDGE offers a [1] http://www.nokia.com (accessed 2003) pretty straightforward business case as a stepping stone [2] http://www.ovum.com/go/ovumcomments/016489.htm before UMTS. For operators with UMTS license a strat- (accessed 2003) egy of EDGE and UMTS co-exist, but for different user segments, can be useful if EDGE is decided to be de- [3] Halonen T, Romero J, Melero J 2002. GSM, GPRS and ployed as widely as (or wider than) UMTS. Another strat- EDGE performance - Evolution towards 3G/UMTS. Wiley. egy can be seen to deploy EDGE as a complementary [4] http://www.ericsson.com/products/white_papers_pdf/ solution, to different coverage areas. This is the most fea- edge_wp_technical.pdf (accessed 2003) sible strategy if there are multi-mode terminals available [5] Penttinen Jyrki. 2002. GPRS in Wireless Data. WSOY. widely. [6] Rantanen J. 2001. The Third Generation Cellular Network The future role of EDGE depends very highly on the suc- Solutions from Operator’s Perspective. Master’s thesis. TKK. cess of EDGE in USA, the availability of EDGE termi- [7] Auramo J. 2002. Enhance Your GSM Networks to 3G with nals, the early success of UMTS, subsidization of UMTS EDGE. Presentation. Nokia. terminals and the completeness of EDGE standardization. [8] www.ericsson.com (accessed 2003) The most probable case is that EDGE will live alongside

UMTS and will in most cases be used as a geographical [9] 3GSM World Congress 2003, Cannes, France, 17-22 Feb- complement to UMTS. The UMTS early success is not ruary 2003. very likely, so it will give operators a chance to consider [10] “Advanced Mobile Networks in Practice” Conference, EDGE as well (which is already ready as a technology) Helsinki 11-14 November 2002 . and invest still to their GSM networks. Also as the tech- [11] Sources within Telecommunications Consultancy company nology develops in the future, it will be more likely that Omnitele Ltd.

15 UMTS and Its Market Analysis

GUOYOU HE TELECOMMUNICATION SOFTWARE AND MULTIMEDIA LABORATORY HELSINKI UNIVERSITY OF TECHNOLOGY FIN-02015 HUT, ESPOO, FINLAND [email protected]

ABSTRACT - Global System for Mobile Communications (GSM) was firstly opened in Finland in 1991. Since the analog cellular systems involved in our life, mobile communications have evolved to its third generation - Digital AMPS (D-AMPS) started its commercial opera- (3G). The richness of features and functionalities with high tion in US in 1994. quality of service in 3G will bring people to a fascinating - Personal Digital Cellular (PDC) was put into commer- world. UMTS is the European vision of 3G mobile com- cial use by NTT in Japan in 1994. munication systems. One of the key functionalities of UMTS is the ability to provide services anywhere and - Code Division Multiple Access (CDMA) started its com- anytime. In UMTS the mobile equipment will be used for mercial operation in Hong Kong and Korea in 1995. any possible purpose such as communication, entertain- Today, multiple 1G and 2G standards are used in world- ment, business and all kinds of services. This essay briefly wide mobile systems, and most of them are incompatible. reviews the evolution of UMTS systems as a 3G platform The most successful implementation of 2G is GSM. Due for mobile communications, classifies the end user ser- to the regional nature of 2G mobile communication sys- vices provided by UMTS, analyses the UMTS markets, tems specifications, GSM did not succeed completely in and presents UMTS vendors’ products and strategies as implementing globalization. well as their recent activities in 3G. The specification on 3G is an evolving process. Many features and Based on GSM, the third generation, 3G, aims to imple- functionalities of UMTS are still under development. ment the globalization of mobile communications. The Vendors battle on pushing their UMTS networks and tech- research for 3G started in 1991. The primary requirements nologies. Though the process of development and deploy- for 3G as described in (Kaaranen, et al. 2001, p. 2) are: ment UMTS will be tough, UMTS shall finally change - The system must be fully specified and world-widely the way of our life and bring people to a brilliant new valid, the major interfaces of the system should be stan- world. dardized and open. - The system must have clearly added value to GSM in all 1 INTRODUCTION aspects and be backward compatible at least with GSM and Integrated Services Digital Network (ISDN) at the Since the introduction of commercial cellular systems in beginning. the late 1970s and early 1980s, mobile communication is - The system must support multimedia and all of its com- evolving to its third generation, 3G. The first generation, ponents. 1G, mobile communication systems transmit only analog voice information and provide basic mobility. The most - The radio access of 3G must provide wideband capacity prominent 1G systems are Advanced Mobile Phone Sys- be generic enough to be world-widely available. tem (AMPS), Nordic Mobile Telephone (NMT), and To- - The services must be independent from radio access tech- tal Access Communication System (TACS). They were nology and the network infrastructure must not limit the incompatible due to the scope of national specifications. services to be generated. The development of the second generation, 2G, mobile With the evolution of communications technologies, the communication systems was driven by the growth need traditional telecommunications and the Internet are merg- for systems compatibility, capacity, coverage and im- ing rapidly. The combination of these two worlds and the proved transmission quality. The development of 2G trends of telecommunications moving to “All IP” require mobile communication systems started in early 1980s. 2G 3G to fulfill more requirements except above primary ones emphasized on the mobile networks compatibility. Speech to fit the changes. transmission was still the main supported services, but data transmissions and supplementary service such as Universal Mobile Telecommunications System (UMTS) fraud prevention and encrypting of user data became stan- (Kaaranen et al. 2003) is the European vision of 3G mo- dard features of 2G systems. The main 2G systems in- bile communication systems. It represents an evolution clude: in terms of services and data speeds from today’s 2G

16 mobile networks. UMTS represents the move into 3G of the first rich call in an end-to-end All IP mobile network. mobile networks. It addresses the growing demand of In 2002, many of the main UMTS vendors announced mobile and Internet applications for new capacity in their progresses in the battle of pushing their 3G networks today’s overcrowded mobile communications. UMTS and technologies (UMTS Forum 2003; UMTS World increases transmission speed up to 2 Mbps per mobile 2003). The main purpose of this paper is to review the user and establishes a global roaming standard. It allows evolution of UMTS, investigates the UMTS services and many more applications to be introduced to a worldwide market situation of UMTS deployment, and analyzes the base of users and provides a vital link between current current situation of UMTS development. multiple GSM systems and the ultimate single worldwide standard systems for all mobile telecommunications. The specifications of UMTS are under development in 2 EVOLUTION OF UMTS Third-Generation Partnership Project (3GPP) (3GPP To address the globalization issues, 3G introduces 2003). To reach global acceptance, 3GPP is introducing WCDMA (Holma and Toskala 2001) as the new radio UMTS in phases: access method. WCDMA is a global system for 3G mo- 3GPP R99 Most of the specifications were frozen in bile communications and allows all 3G subscribers to be March of 2000. It laid the foundations for high-speed traf- able to access all 3G networks. It has better spectral effi- fic transfer in both circuit switched and packet switched ciency than Time Division Multiple Access (TDMA) in modes by defining enhancements and transitions for ex- certain condition and is more suitable for packet transfer isting GSM networks and specifying the development of than TDMA based radio access. For using WCDMA, new new radio access network. radio access network, UTRAN, composed of Base Sta- tion (BS) and Radio Network Controller (RNC), has to 3GPP R4 Most of the core technical specifications were be added due to the incompatibility between WCDMA frozen in March 2001. It is a minor release with the evo- elements and GSM equipment, and the interoperability lutions including Universal Terrestrial Radio Access Net- of GSM/UMTS has to be handled. For taking care of the work (UTRAN) access with Quality of Service (QoS) en- interoperability, EDGE Radio Access Network (E-RAN) hancement, Circuit Switched (CS) domain evolution with is modified to be able to broadcast system information introducing Mobile Switching Center (MSC) server and about WCDMA radio network in its downlink and inter- Media Gateways (MGWs) based on IP protocols, enhance- working functionality is introduced into the evolved 2G ments in Location Communication System (LCS), Multi- MSC/VLR for handling WCDMA. media Messaging Service (MMS), Mobile Execution En- vironment (MexE), etc. The 3GPP R99 (3GPP 2002a) of UMTS introduces WCDMA as radio access and effectively utilizes existing 3GPP R5 Most of the specifications and technical reports GSM/GPRS system providing the basic communication were frozen in March 2002 or June 2002. It is a major services for both (Circuit Switched) CS and Packet release aiming to utilize IP networking as much as pos- Switched (PS) traffic together with a rich set of Value sible. IP and overlying protocols will be used in both net- Added Service (VAS) and supplementary services. The works control and user data flows, i.e. implement “All Core Network (CN) is divided into CS domain and PS IP” network, but the IP-based network should still sup- domain for handling circuit switched and packet switched port circuit switched networks. The features of this re- traffic respectively. lease mainly include the introduction of IP Multimedia Subsystem (IMS) (3GPP 2002e), enhancement in The CS domain contains 3G MSC/VLR and Gateway Wideband Code Division Mutiple Access (WCDMA) MSC (GMSC). These two elements can be physically (Holma and Toskala 2001), MMS, and LCS. In 3GPP R99 separated or combined. The 3G MSC/VLR evolves from the basis for the UMTS radio access is WCDMA. In 3GPP GMS MSC/VLR by merging the transcoders required for R4/R5 GSM/EDGE Radio Access Network (GERAN) is speech coding conversion from the radio network to MSC/ specified as an alternative for radio access to build a VLR. The Visitor Location Register (VLR) is an integral UMTS mobile network. part of the Mobile Switching Center (MSC) in 3G. The 3G MSC/VLR is responsible for CS connection manage- 3GPP R6 It is still being defined with the target June 2003. ment activities, Mobility Management (MM) related is- In this release, a lot of enhancements and improvements sues such as location update, location registration, pag- in IMS, Multimedia Broadcast/Multicast Service ing and security activities. The GMSC takes care of the (MBMS), MMS, QoS, GERAN will be specified. Many incoming/outgoing connections to/from the external net- new services such as digital rights management, speech works. It initiates a location info retrieval procedure to recognition and speech enabled services and priority ser- find the correct 3G MSC/VLR for call path connection, vice will be specified. and establishes a call path towards the 3G MSC/VLR UMTS is already a reality. Japan launched the world’s under which the addressed subscriber is to be found. first commercial WCDMA network in 2001. Nokia and The PS domain is evolved from General Packet Radio AT&T Wireless complete first live 3G EDGE call on Service (GPRS) (GSM Association 2003). It contains November 1, 2001. Telenor launched the first commer- Serving GPRS Support Node (SGSN) and Gateway GPRS cial UMTS network in Norway in December 1, 2001. On Support Node (GGSN). The SGSN node supports packet February 20, 2002, Nokia and Omnital Vodafone made communication towards the access networks for both GSM

17 Base Station Subsystem (BSS) and UTRAN. SGSN switches. The MSC server contains CM main functional- mainly takes care of Mobility Management (MM) related ity and takes care of MM. VLR is also integrated into it. issues such as route update, location registration, packet The MGW contains the functionality of performing ac- paging and security. The GGSN node maintains the tual switching and network inter-working. It may contain connections towards external packet switched networks other functionality such as performing circuit packet con- such as Internet. This node is responsible for route info version in VoIP calls, etc. The relationship between MSC/ retrieval and routing packets to/from SGSN for further GMSC server and MGW is one to multiple. It means that relaying. It also takes care of session management. one MSC/GMSC server can control numerous MGWs. The Registers part is composed of Equipment Identity The number of MGWs under one MSC/GMSC server is Center (EIR), Home Location Register (HLR) and Au- scalable and the MSC server amount may be dimensioned thentication Center (AuC). This part does not deliver traf- in the system. fic. Instead it contains addressing and identity informa- The largest new functionality in 3GPP R5 (3GPP 2002d) tion required for MM and security for both CS and PS. is IP Multimedia Subsystem (IMS) (3GPP 2002e). IMS HLR contains permanent data of the subscribers and is has a uniform way to Voice over IP (VoIP) and other real- responsible for MM related procedures. AuC is a data- time and non real-time IP services such as multimedia base generating the Authentication Vectors that contain services. All the access networks can be IP based. The the security parameters used for security activities, it can traffic can be always packet switched, and all the services be an integrated part of HLR. EIR contains the identifica- can be moved to the PS domain. The HLR is evolved to tion information related to the User Equipment (UE). Home Subscriber Server (HSS) providing enhanced fea- Compared to R99, the changes in 3GPP R4 (3GPP 2002c) tures for support IMS. 3GPP R5 contains all the possi- are extended remarkably to CN instead of in the radio bilities for traffic treatment. No matter the traffic coming access network. Especially in the CN CS domain, the from the access network is packet switched or circuit MSC/VLR and GMSC are evolved into (G)MSC server switched, which can be relayed to the external network and MGW to separate Communication Management (CM) either in circuit switched or in packet switched manner. and actual switching as well as related functions into sepa- In 3GPP R5, the GERAN can be connected to the CN rate physical entities. with Iu interface. Regarding to this interface, the traffic The MSC/GMSC server is evolved from the MSC/GMSC. from GERAN can get the same treatment as the traffic It mainly comprises the call control and mobility control from the UTRAN. When IMS is in use, the CS domain parts of a MSC/GMSC. Whole connection process is con- will not be need any more. So one of main differences trolled by the (G)MSC server(s), user data goes through between R4 and R5 is that CS can quit service in R5, and MGWs, which maintain the connection and act as the whole network will finally transfer to “All IP”.

Category Description Fun WWW, video, post card, snapshots, text, picture and multimedia messaging, datacast, personalisation applications (ring tone, screen saver, desk top), jukebox, virtual companion, etc. Work Rich call with image and data stream, IP telephony, B2B ordering and logistics, information exchange, personal information manager, dairy, scheduler, note pad, 2-way video conferencing, directory services, travel assistance, work group, telepresence, FTP, instant voicemail, colour fax, etc. Media Push newspaper and magazines, advertising, etc. Shopping E-commerce, e-cash, e-wallet, credit card, telebanking, automatic transaction, auction, micro-billing shopping, etc. Entertainment News, stock market, sports, games, lottery, gambling, music, video, concerts, adult content, etc. Education Online libraries, search engines, remote attendance, field research, etc. Peace of mind Remote surveillance, location tracking, emergency use, etc. Health Telemedicine, remote diagnose and heath monitoring, etc. Automation Home automation, traffic telematics, machine-machine communication, etc. Travel Location sensitive information and guidance, e-tour, location awareness, time tables, e-ticketing, etc. Add-on TV, radio, PC, access to remote computer, MP3 player, camera, video camera, watch, pager, GPS, remote control unit, etc.

Table 3-1: The possible types of services in 3G networks

18 Service Applications Users Revenues category (2010) Mobile • Messaging (E-mail), Travel assistance Mobile office, 15% Intranet/ (WWW) Business user Extranet Access • Mobile sales, Technical services • Teleworking, Access to corporate database • Video telephony, Conferencing • Fleet management, Warehaouse Mobile Internet • Messaging (E-mail, SMS, MMS) Business user, 3% Access • Download video, music, streaming Consumer • VoIP, Video over IP • m-banking • m-commerce (m-purchasing), trading • www travel • www Infoservices Customized • Information (photo, video, music Business user, 28% Infotainment download) Consumer • www travel • Education (schools, universities) • Mobile messaging, Chatting (SMS, MMS) • Gaming • m-shopping, banking, e-wallet, micro- payment Multimedia • Extension of SMS User 15% Messaging • MMS: Image, Video, Unified messaging, Mobile postcard, video/audio clip • MS Office document • Mobile chatting • Machine to machine communications • Photo messaging • Music • Video messaging Location-based • Navigation (person car) User 3% Services • Localized Info (yellow pages) • Location-based m-commerce • Telematics • Trading (vehicle, goods, person) Rich Voice • Telephony/Conferencing User 34% • Video-telephony, conferencing, presence • Telemedicine • Teleworking (building industry etc.) • Multimedia communication (IMS)

Table 3-2: UMTS service categories and their applications

3GPP R6 is being specified currently. Based on R5, it service(s); content provider is a party creating and pro- will enhance and improve the existing functionalities and viding the end user services. In real life, one or more par- services in the fields including Open Service Architec- ties could be the same commercial party, for example the ture (OSA), QoS LCS, MMS, security, GERAN and ra- service provider and carrier provider are often the same dio interface, etc. The main implementation in this phase company and also acts as service provider to some ex- will be IMS improvement and the interworking between tend. Wireless Local Area Network (WLAN) and UMTS. R6 With the development of 3G, more and more applications will also specifies MBMS, speech recognition and speech are emerging for UMTS. There might never be a single enabled services. killer application to monopolize the application markets. It is predictable that multiple and rich applications will coexist in the UMTS systems. The applications of 3G can 3 END USER SERVICES be divided into different categories based on different The business or value chain in UMTS consists of end user, criteria by different organizations and vendors. carrier provider, service provider and content provider. UMTS Forum (2003) divides the near-term 3G data ser- Carrier provider is the party maintaining both access net- vices into content connectivity and mobility, which are work and CN; service provider is the one providing subdivided into six categories of services as shown in

19 Vendors Market shares Ericsson 33 % Nokia 32 % Siemens (NEC) 15 % Nortel 8 % NEC (Siemens) 4 % Alcatel 4 % Lucent 3 % Motorola 1 %

Table 4-1: Estimation of UMTS (WCDMA) sales volume market share (late 2001) (UMTS World 2003)

Table 3-2. From the prediction of UMTS Forum, the big- here, we can have an overall image for future UMTS gest revenues of the operators will be from Rich Voce, market. Customized Infotainment, Multimedia Massaging and Mobile Intranet/Extranet Access. 4.1 3G Market Shares No matter how the applications will be classified and what criteria will be used for the classification, the bottom line Table 4-1 shows the estimation of WCDMA market shares is that computer and Internet applications will be merged based public information about UMTS contracts until late with communication and location based services. Many 2001. The actual situation may vary a lot. Anyway we can networks operators, vendors and third party service pro- see the rough position of different players in the battle of viders have been already creating many applications and pushing their UMTS products. demos. Though it is difficult to enumerate the concrete Table 4-2 illustrates UMTS world’s estimation on 3G services in 3G networks, the possible services that will be markets in different geographical areas. The vendors will available in 3G networks are listed in Table 3-1. battle for these markets for pushing their products and technologies. Table 4-3 and 4-4 present the biggest GSM operators and 4 UMTS MARKET mobile markets by number of subscribers respectively until This section presents the estimated market shares and 2002. These operators and areas are the potential inves- predications for 3G and mobile communications based tors and places for deploying UMTS first in the next sev- on publicly available information. From the information eral years.

Geographical Area Market Size Asia / Pacific ~ 40% Europe (East & West) ~ 30% Americas (North & South) ~ 30% Africa / Near East ~ several %

Table4-2: 3G (cdma2000 and WCDMA) market sizes around 2005 (UMTS World 2003)

GSM Operator Subscribers in millions (June '02) China Mobile 123 Vodafone Group 119 T-Mobile 56 China Unicom 52 Orange 41 AT&T Wireless 27 Telecom Italia Mobile 27 Singtel 24 Cingular 22 Telefonica Moviles 19 mmO2 18 Turkcell 14 Note: These are carriers that operate GSM networks and have GSM customers, but these numbers represent the entire customer base, and not just GSM subscribers.

Table 4-3: Top 12 GSM Operators By Subscribers (June 2002) (UMTS World 2003)

20 4.2 Vendors Products and Strategies WCDMA network infrastructure to Tele2/Tango in Lux- embourg and Liechstenstein. The latest activities of Currently, many vendors have involved in UMTS systems Ericsson are that it demonstrated mobile video call in a development and deployment. The largest UMTS ven- fully integrated dual-mode WCDMA/GMS platform, and dors are listed in Table 4-1. Currently it seems that no was going to provide Orange UMTS core networks equip- network vendor can supply all equipment and components ments and associated integrated services. to the full 3G networks. Quite a few of them are the main contractors to build 3G networks. Normally network ven- Nokia provides the whole systems from terminals and base dors bring in partners in different areas such as services stations to core network solutions for GSM, GPRS and and applications providers, handset manufacturers, etc. UMTS. The products spread in a wide range with various Since late 2002, the development and introduction of models. Nokia also provides all kinds of platforms for UMTS are more and more active though the economic 3G systems. situation is still dim. Many vendors announced launching Nokia is also very active in 3G development and deploy- of their UMTS products, and more operators introduced ment. Nokia announced to participate in 3G expansion UMTS operations in reality. Following presents the main with J-PHONE in Japan and to supply WCDMA 3G net- vendors’ UMTS products and their 3G related activities work to CHT Taiwan in October 2002. In November 2002, in last four months according to recent public informa- Nokia announced that it had launched 15 phones support- tion (UMTS Forum 2003; UMTS World 2003). ing MMS and delivered the world’s first GSM/EDGE 3G mobile phone, 6200 tri-band (GSM/GPRS/EDGE 850/ 1800/1900 MHz) phone, for operator controlled live net- Ericsson provides the whole range of 2G and 3G Mobile Systems and end-to-end system elements including infra-

Area Subscribers in millions Until China 200.3 Dec 2002 USA ~137 June 2002 Japan 72.8 Nov 2002 Germany 64.4 July 2002 Italy ~47 Late 2001 United Kingdom ~45 Late 2001 France ~34 Late 2001 South Korea ~32 Late 2001 Spain ~28 Brazil ~27

Table 4-4: Biggest Mobile Market (UMTS World 2003)

structure, terminals, applications and expertise. Who also work testing, and successfully performed 3G WCDMA provides total solutions from systems and applications to call handover to commercial GSM network with Vodafone services and core technology for mobile handsets. With in Italy. In December 2002, Nokia successfully demon- Sony Ericsson, Ericsson also provides complete mobile strated IP mobility services using Nokia’s IMS, which is multi-media products. expected to have the first commercial release in 2003, followed by IMS capable terminals in 2004. Nokia also Ericsson is very active in 3G development and deploy- demonstrated Service Area Identity (SAI) positioning ment. In October 2002, Ericsson announced to partici- technology in Radiolinja’s pre-commercial WCDMA net- pate 3G expansion with J-PHONE in Japan, to jointly work, and agreed to deliver WCDMA 3G Network to develop and supply WLAN access solutions for 2G and Taiwan Cellular Corp. in December 2002. Until Febru- 3G networks with Agere and Promix, and to supply ary 2003, Nokia has delivered 10,000 Nokia 6650 phones CDMA2000 1xEX-DO overlay system to Vesper in Bra- globally to operators for testing. The commercial deliv- zil. In the same month, Ericsson also confirmed 10000 eries are expected to start first half of 2003. UMTS/WCDMA macro base stations were shipped worldwide. In December 2002, Ericsson and AT&T Wireless Siemens in partnership with NEC provides UMTS radio completed the first WCDMA/UMTS call in a live net- solution (FDD and TDD), carrier-class switching for work in the Americas. Ericsson was also chosen as pri- UMTS and many kinds of enabling services for 3G sys- mary 3G supplier by Danish operator TDC Mobil, the tems. delivery planned to start in January 2003. In January 2003, Siemens shipped first U10 3G phones to selected opera- Ericsson agreed to deliver seamless 2G/3G WCDMA tors in October 2002. In December 2002, Siemens agreed networks to Far EasTone in Taiwan, to supply new Video to supply and install all the necessary components of a Gateway System to Hutchison globally, and to provide

21 complete 3G mobile network for P&T Luxembourg in In November 2002, Motorola agreed to introduce Next the second half of 2002. Six weeks after the agreement, Generation Network Management Platform from HP to Siemens made the first live call via UMTS networks with speed up UMTS network deployment. In January 2003, UMTS mobile phone (U10) at P&T Luxembourg Motorola unveiled mobile handsets, applications for 2003 including A835 powered with 3G technology in Shang- Nortel offers end-to-end UMTS network solution includ- hai. In February 2003, Motorola announced the industry’s ing terminals, radio access networks and core networks first live over air call of its fully integrated wireless hand- for UMTS systems. set platform, which supplies handset designers and manu- In Q4 2002, Nortel demonstrated the world’s first UMTS facturers with a complete silicon-to-software platform call using IP-based UTRAN, agreed to carry out a pro- solution. gram to realize the first 3G UMTS trial test with Samsung NEC offers products in the fields of mobile communica- and Vodafone in Spain, and announced the ability to pro- tions and optical networks. vide integrated solutions for seamlessly link Wireless Wide Area Network (WWANs) with WLAN. Nortel was se- In October 2002, Hutchison Whampoa ordered two mil- lected to supply CDMA2000 1xEV-DO networks by PT lion 3G video devices (handsets) from NEC. In February Wireless Indonesia (WIN) in January 2003. In February 2003, NEC demonstrated its advanced Internet server 2003, Nortel demonstrated its next generation Wireless software at 3GSM World Congress in Canned, France. Data Network solutions featured with live 2.5G and 3G All vendors have different product schedules and devel- wireless services at 3GSM World Congress in Cannes. opment status. It is a big secret for all of the vendors. Alcatel is mobile networks vendor offering second and Based on recent announced activities and achievements, third generation (2G and 3G) solutions from networks, it is obvious that Ericsson and Nokia are the key vendors applications and terminals to implementation and opera- in 3G. Currently, they can provide the most complete so- tion. For 3G networks, Alcatel offers end-to-end solution lutions for 3G/UMTS systems. It seems that Ericsson has that includes networks, applications and services. Together more 3G market shares than Nokia in the competition in with partners, Alcatel is developing a wide array of appli- last four months, but Nokia is also very active in devel- cations and packages that are ready to run. oping and pushing its products and technologies. Alcatel is a new vendor in 3G and is quite active in developing its In Q4 2002, Alcatel launched fully functional 3G/UMTS 3G network solutions and pushing its pushing products to mobile systems in the Netherlands and Germany, and con- the market. The strength of Alcatel is in its 3G service firmed that it had delivered and installed EDGE-equipped solutions. Siemens in partnership with NEC can also pro- mobile infrastructure solutions over 100 networks world- vide whole 3G/UMTS systems to operators now. Nortel’s wide. Austrialia’s m.Net tested Alcatel’s advanced mo- 3G products are mainly characterized in WLAN, IP and bile services platform, which supports SMS, WAP, MMS, ATM solutions for 3G. Its products are still under testing GPRS and 3G applications. In January 2003, Alcatel and trial. Lucent has small shares in the 3G market now. agreed to deliver and install its complete Evolium 3G end- Its 3G solution have not been fully matured yet. NEC in to-end solution including UTRAN, the core network sys- partnership with DoComo is successful in developing its tems (packet and circuit), and associated multimedia 3G handsets. Its development of mobile Internet server equipment and terminals for a 3G/UMTS field trial by solution seems to have reached the final stage. Motorola Etisalat in the United Arab Emirates. In February 2003, has not been very successful in developing its 3G solu- Alcatel chose IPWirelss to work on mobile broadband tion yet, but more its 3G handsets will come to market in access solution based on TD-CDMA (TDD), introduced 2003. the Ultimate broadband wireless access solution for 3G/ UMTS backhaul, and announced to deliver 3G/UMTS In the battle of developing and pushing 3G products, it is multimedia services for Orange France. obvious that all of the vendors get partners in one or more fields to speed up their solutions. Though the current eco- Lucent provides 3G solutions - CDMA and UMTS. Cur- nomic situation is dim, all the vendors are very active in rently, Lucent delivers 3G networks. It has already up- getting progress on their 3G/UMTS solutions and push- graded more than 35,000 base stations to add 3G capa- ing them to market. bilities. In November 2002, Lucent and Option demonstrated successful data calls using jointly developed wireless mo- 4.3 Terminal Availability dem cards for 3G UMTS networks. Lucent and Novatal Except the competition in developing and pushing 3G/ Wireless also unveiled jointly developed wireless modem UMTS networks solutions, the battle in innovation of card for 3G UMTS networks in February 2003. mobile terminals and pushing them to market is even hot- Motorola launched the commercial 3G CDMA 1X ser- ter due to more competitors involved in it. Table 4-5 lists vice with access speeds up to 144 kb/s in Japan In 2002. the main 3G terminals that have already been available or Motorola is developing UMTS systems based on its to be available in near future based on public informa- CDMA technology. tion.

22 5 CONCLUSION deployment of telecom products and technologies are acceleration processes for both vendors and operators. It Mobile communications move to its third generation has requires the vendors have to output much more complex become a must trend. The specifications for 3G have been products and technologies within much shorter lifecycle, evolving from R99 to R6. Most of the technical specifi- and the GSM operators invest in new systems or update cations for R5 were frozen until June 2002. R6 is under their existing systems though they were just put in opera- defining with the target June 2003, but the current esti- tion. It’s a contradiction process demanding both the ven- mated time for finalizing R6 is December 2003. 3G/ dors and operators have to have powerful financial back- UMTS specifications reveal unimaginable splendid life ing to get around it. world to all kinds of consumers, they intend to make everything available on moving, no matter when, where, who UMTS provides rich end user services. Among these ser- and what. vices, Rich Voice, Customized Infotainment, Multimedia Massaging and Mobile Intranet/Extranet Access will bring In the 3G/UMTS development aspect, many vendors have the biggest revenues for the operators in near future. involved in the venture actively. Most of them have already launched their 3G/UMTS products and push them On the market side, the most prospective potential mar- to real operation or put them under trial. The develop- kets for 3G/UMTS are still China, Asia Pacific Area, USA ment and deployment of 3G/UMTS is also an evolvement and Europe. Compared to PSDN and GSM, 3G/UMTS process, it is estimated that only the key 3G vendors have systems are luxury systems for most of the ordinary us- been developing corresponding to R4/R5 products or have ers, especially at the beginning. This luxury somehow lim- already had R4/R5 products partly though all of the ven- its the deployment of 3G/UMTS systems in economic dors are battling against pushing their products and tech- undeveloped regions though these places might still be nologies. Looking back the path for telecommunications, virgin for mobile communications. To get around it, the it’s not difficult to find that both the development and

Vendor Product Key Features Key Functionalities Availability LG UMTS 1) High Speed Data Transmission 1) IDual Browser, W IM 3Q.'02 Handset 2) M ultimedia Player (Built-in Cam era, 2.2" 2) K-Java TFD Color LCD, M PEG4 for Video 3) USIM Streaming/Download) 4) Plug and Play Application 3) User Friendly User Interface(Easy Color 5) M P3 GUI) 6) Bluetooth, USB Connectivity 7) Position Location 8) Voice Recognition Motorola A820 1) Multi-functional voice/data 1) Customizing messages with 2002 embedded MP3 player video or audio files 2) Integrated video camera 2) MP3 3) M ulti-call, multi-task 3) MMS 4) Support 2G, 2.5G and 3G

A835 GPRS/EDGE/WCDMA EMS and MMS, Java(TM ) 2003 NTT FOMA 1) High speed packet data transmission 1) 3G handset compatible with the 2002/2003 DoCoMo series (receiving speeds of up to 384Kbps) FOMA network in Japan NEC 2) Circuit switched data transmission speed 2) Download and play various up to 64kbps video files such as music video, 3) Multitask feature – performs up to 3 movie previews, news and sport. activities simultaneously; voice call, use of i- 3) i-mode. mode and a terminal function such as 4) Video clip scheduler, calculator, address book etc 4) External connector USB interface NEC FOMA Similar features as above Similar functionalities as above 2002 N2002 Nokia 6650 1) Highe speed data: up to 57.6 kps in CS 1)Wireless Connectivity: Commercial data networks and 384 kps (downlink) and 64 Infrared, Bluetooth, Wireless phone deliveries in kps (uplink) connection to a compatible PC first half 2003 2) Built-in camera for shooting video at over 2) Send/receive pictures, video ten frames per second, a 4096-color display, clips, graphics, play games and MMS capability for sending and 3) M essaging: Combine receiving clips, video/picture, text, and voice into 3) Works both in GSM 900/1800 networks multimedia messages. Send and in the ne WCDMA networks. multimedia messages to a 4) In the W CDM A network, talk, snap, and compatible phone or PC. send pictures at the same time. Concatenated text messages, picture 5) Voice dialing, Voice recorder, Integrated messaging handsfree speaker 4) WAP: WAP 1.2.1 over GPRS data in both GSM and W CDM A mode

6200 1) Tri-band (GSM /GPRS/EDGE/850/ 1) MMS Commercial 1800/1900MHz) 2) XHTML deliveries Q1 2) Data rate up to 118kb/s 3) Java(TM) 2003 Siemens U10 1) Color screen 1) Wap 2.0 21/10/2002 2) W ork in both in both W CDM A networks 2) MMS, Video clip and e-mail and in GSM/GPRS 900/1800/1900 frequency 3) Bluetooth, USB or infrared bands connectivity 3) Integrated camera. 4) MP3 4) Data stream up to 384 kb/s Table 4-5: 3G Terminals

23 vendors have to provide cheap UMTS terminals, and the 3GPP, 2002c. TS 23.002 Network Architecture, Release 4, operators have to offer cheap 3G services. V4.6.0 The aim of 3G is to evolve to “All IP”. To reach this tar- 3GPP, 2002d. TS 23.002 Network Architecture, Release 5, get, both vendors and operators have quite long to go V5.6.0 ahead. The vendors have to overcome all the critical tech- 3GPP, 2002e. TS 23.228 IP Multimedia Subsystem (IMS), Stage nical problems such as delay, echo and other QoS related 2, Release 5, V5.7.0 issues for real time communications. The operators have 3GPP, 2003. http://www.3gpp.org (20 Jan 2003) to expand the coverage of 3G/UMTS systems with great financial support. GSM Association, 2003. What is General Packet Radio Ser- vice? [online]. http://www.gsmworld.com/technology/gprs/ intro.shtml (11 Jan 2003) REFERENCES Holma Harri, Toskala Antti, 2001. WCDMA for UMTs: Radio Access for Third Generation Mobile Communications. Chichester: John Wily & Sons, Ltd. 3GPP, 2002a. TS 23.002 Network Architecture, Release 1999, Kaaranen Heikki, etc., 2001. UMTS Networks Architecture, V3.6.0 Mobility and Services. England: John Wiley & Sons, Ltd. 3GPP, 2002b. 3GPP Specifications - Release contents and func- UMTS Forum, 2003. http://www.umtsforum.org/servlet/dycon/ tionality (online). http://www.3gpp.org/specs/releases- ztumts/umts/Live/en/umts/Home (20 Feb 2003) contents.htm (15 Jan 2003). UMTS World, 2003. http://www.umtsworld.com (13 Jan 2003)

24 UMTS Investment Study

YAOJIN YANG HELSINKI UNIVERSITY OF TECHNOLOGY [email protected]

1 INTRODUCTION 1.1.2 Determinates of UMTS investment Determinates of UMTS investment are various. From tech- 1.1 Background nical viewpoint, for example, the specific technical features of UMTS limit the scalability of the network infra- The UMTS investment study has been performed for the structure compared to the GSM and intermediate plat- telecommunication business seminar of HUT (Helsinki forms such as GPRS or EDGE. So key aspects of the University of Technology). The study didn’t cover all UMTS network need to be designed for its full capacity perspectives which are concerned by the rollout of UMTS. in the initial stage. From economic viewpoint, compared However, it only bases on the investment perspective from to GSM, the UMTS network requires higher investment operators’ side. The report was created in terms of this early in the network rollout. The investment decisions are study. complicated by the factors including the huge license cost, rollout obligation agreed with government, and the tech- 1.1.1 UMTS background nological and market uncertainty. UMTS (3G), Universal Mobile Telecommunications Sys- tem, is an evolution in terms of services and data speeds 1.1.3 Model of UMTS market from today’s mobile networks (2G and 2.5G). It is the natural evolutionary choice for operators of GSM networks, currently representing a customer base of more than 747 million end users in over 180 countries and representing over 70% of today’s digital wireless market [1]. UMTS is synonymous with a choice of WCDMA radio access technology that has already been selected by 112 licensees worldwide by using fresh radio spectrum. To- day, it is already a reality. In some countries the pre-commercial networks have launched; In Japan, even the commercial network has opened. However, from the investment point of view the rollout of UMTS network is still complicated by several factors. Figure1. the model of UMTS market [2] First, as a packet-based network, UMTS will have different technical characteristics than GSM networks. Among The technical module represents key technological rela- other things, these affect the optimal location of anten- tionships. Technology, coverage, capacity and data rate nas, the way capacity is built, and quality of service is- jointly decide the size, availability and quality of the sues. Second, the high investment cost requires the rapid UMTS network. These features of the network in turn expansion of market share and the generation of revenues. determine the provided services. The services will strongly Third, the sunk costs of 3G licenses and networks create influence the number of users. additional financial stress. Fourth, the regulatory environment for 3G services is somewhat in flux, increasing un- The investment module captures the relations that shape certainty for investors. Last but not least, the rollout of the decisions of a market player with regard to the level UMTS infrastructure capacity and services is likely af- and structure of new investment. The factors, such as the flicted with strong positive feedback effects. In such situ- planned services and the expected cost and revenue ations, a sufficient revenue stream is dependent on quickly streams, jointly determining expected profitability, related reaching a critical mass of subscribers. How rapidly con- to an investment will be important in this module. The sumers accept UMTS services, in turn depends on the number of competitors (as well as the demand and supply availability of infrastructure and services [2]. conditions) will affect the expected revenues as more intense competition, other things equal, will also be taken into account in the module. Moreover, regulatory obliga-

25 tions, such as rollout requirements, will be part of the Above figure shows a cyclic chain that represents a gen- module as well. eral UMTS business model from operators’ point of view. In an ideal case, each node in the chain is supported by The supply side module depicts the main relations on the previous node. For example, investment supports service supply side of the market. At its core is the insight that the providing and coverage expansion, service and coverage total cost in every period is the sum of the fixed and vari- support demand increasing, demand supports profit gen- able costs, whose determinants are represented in the eration, and profit in turn supports further investment. module. A key driver of fixed cost is the investment cost, Through the chain the UMTS business get advanced. which is directly related to the planned network expansion, the equipment costs, and the license fee. Variable However, the reality of the combination of huge UMTS costs are defined as the costs in any period varying with infrastructure costs and current negative financial situa- the output level. Key drivers are the total output of the tion is deconstructing the base of the investment. Conse- service provider, such as number of customers and num- quently, few services and less coverage are provided, de- ber of minutes, as well as the resource costs, such as la- mand is not increased, profit is not as expected. Finally, bor and materials. Variable costs influence retail and further investment can not be achieved and the whole in- wholesale supply prices. Fixed costs influence these prices dustrial is facing a serious risk. only indirectly as firms well attempt to impost a mark-up on the incremental costs. 1.3 Structure of the report The determinants of wholesale and retail demand are represented in the demand side module. Income, the price of Chapter 1 is an introduction about the background knowl- substitutes, the price of complements, availability of con- edge of the study, and the UMTS and its relative areas. tent, and the quality of service are the most important Also a problem model of UMTS business from opera- variables. tors’ viewpoint is given in the same chapter. Chapter 2 describers the goals and methodology of the study. The Demand and supply-side factors are integrated in the com- results of the study are explained in chapter 3. In the last petition module. This module shows the factors determin- chapter a conclusion is drawn. ing the actual market price, market demand, and the resulting wholesale and retails revenues. For example, the more competitors operate in the wholesale market, the 2 GOALS AND METHODOLOGY low wholesale prices. High wholesale prices would increase revenues, but reduce sales volume. 2.1 Scope of the study 1.1.4 Predictions of the UMTS development The scope of the study focuses on the UMTS investment issues from operators’ point of view; it includes problem Nortel Networks predicts that UMTS will develop in four model, combined costs and cost saving, revenue genera- stages. Pre-commercial launches take place in the end of tion, and strategies. 2002. In 2003 UMTS will be rolled out in significant cities. There will be considerable take-up in 2004 and by 2005 30-35% of terminals sold in Europe will be UMTS. 2.2 Goals of the study So far, the UMTS reality fits the prediction. For example, 1. Create and analyze problem model from operators’ point Vodafone launched pre-commercial UMTS services in the of view. Isle of Man, Monaco in 2002, and Sonera, a Finnish operate, has announced in 2003 the commercial UMTS ser- 2. Analyze combined costs for operators. vices will take place in four significant Finnish cities: 3. Analyze possible cost saving method. Helsinki, Tampere, Turku and Oulu. 4. Discuss possible revenue sources for operators.

1.2 Problem model 5. Analyze strategies for operators’ business success.

2.3 Methodology The study based on the literatures collected from Internet. The general study process consists of literature search- ing, reading, analyzing, presentation preparing and report writing. Also several interviews with people from industrial have been conducted during the study.8

Figure 2. The general UMTS business model from operator’s viewpoint

26 3 Results

3.1 Costs The success of UMTS business very much depends on customers’ willingness to pay for new services, and the quality and availability of services. As explained in the Table1: License Fees per Capita. Source: (Duarte chapter 1, providing services needs the investment from 2001) and (OECD 2001) operators. The costs that will be first incurred are for li- The high license fees are beginning to have consequences censes and infrastructure. Following this investment, op- for operators and the financial situation of wireless com- erators will incur additional costs for developing applica- panies. The combined effect of sunk license fees and a tions and attracting customers to new services. slower growth in the telecom industry leaves operators in The combined costs for operators can be broken into 8 a precarious position. The increased uncertainty of the categories: license cost, operational cost, network set up, market would recommend postponing UMTS infrastruc- content acquisition, product development, customer ac- ture investment until more information about the future quisition, handset subsidization, and fixed marketing costs developments is revealed. However, the license payment [2]. increases the pressure on operators to generate revenue quickly, which would require the rollout of the UMTS infrastructure. Thus, cost saving becomes an important issue. For example, infrastructure sharing would be a ra- tional strategy for operators to minimize the risk.

3.1.2 Infrastructure cost The costs of 3G infrastructure will depend partly on the cost of capital, but it is also influenced by a wide variety of other factors, such as site costs, microwave links, back haul, base stations, etc. The costs of infrastructure will also depend on the cost of equipment. Several equipment vendors in mobile telecommunications are currently also facing financial difficulties and will be relying on revenues from sales of new equipment to improve their posi- Figure3: Breakdown of Costs for 3G Infrastructure and tion. This however is offset by the powerful position of Service. Source: J.P. Morgan/Arthur Anderson (2001) some operators given recent and future consolidation com- and [2] bined with the fierce competition in the mobile equipment sector [3]. The financial problems of equipment vendors could add These cost estimates imply the time in which costs will to the already difficult position of operators when it comes be incurred. Costs incurred during the early stage include to financing the cost of equipment, which is often pro- the product development, license and network setup costs. vided by vendors through either their own assets or through Handset special arrangements with banks. The position of opera- subsidization, customer and content acquisition will come tors in negotiations with vendors is influenced by the size later when services start to be deployed. Costs categories of the order, however it would be difficult for an operator unique to the 3G environment include the license, and particularly in the current market conditions to turn down content and customer acquisition. the lowest cost alternative for equipment [2].

3.1.1 License cost 3.1.3 Infrastructure sharing The UMTS license fee was determined both by the The UMTS rollout requires a very heavy investment from method, with which governments distributed the, the num- operators, so cost saving is important to migrate possible ber of licenses issued, as well as the participants’ busi- risks. Infrastructure sharing has been proposed as the key ness plans and willingness to pay. The high level of the way to reduce costs. Four main network-sharing options bids, particularly those in the earlier European 3G auc- are open to UMTS operators. These are site sharing, tions, surprised both governments and participates. Some UTRAN sharing, common-shared networks and geo- samples from Europe are as follows. graphically split networks. The site-sharing option is now widely deployed across Europe. Operators typically share both the base-station location and masts for antenna placement. Power equip-

27 ment, transmission equipment and antennas may also be shared, but operators ultimately deploy their own radio base stations. Site sharing can be done on a site-by-site basis or as part of a coordinated rollout plan, which can lead to further savings in network planning, civil works and operations costs. The sharing arrangement can be made directly between the operators, or may include a tower company or other third party. Site sharing is ben- Table2: Potential Savings through Infrastructure eficial in a wide range of scenarios. In densely populated Sharing. Source: Lehman Brothers, 2001, Wireless areas where base-station sites are scarce and expensive, Services Industrial Overview site sharing can reduce expenditure, while in rural areas transmission and power costs can be reduced. Regulators and other government agencies are usually keen to pro- 3.2 Revenue mote site sharing because it reduces the environmental impact of antennas and enables new operators to build To avoid the risks derived from the high investment in their networks more easily by reusing existing sites. UMTS business, it not only requires cost saving but also requires revenue generation. The ability of generating The second option involves the sharing of the UMTS ra- revenue is a core factor that determines if the business dio-access network (RAN), which is usually referred to will be successful. Analyzing the potential revenues for as the universal terrestrial radio-access network UMTS services is a challenging, since it is as yet unknown (UTRAN). In this model, operators share all the elements what types of applications will drive demand for UMTS of the UTRAN including base stations, radio network based services and is also unknown how many of the les- controllers (RNCs) and transmission facilities. Power sons learned with GSM can be applied to UMTS. amplifiers and antennas could also be shared. Within the shared network, each operator employs its own licensed Comparing UMTS with GSM the first difference one spectrum and deploys cells with individual control and observes is the state of the diffusion of mobile telephony network management. Operators also deploy their own when the upgrade from analog services to GSM services separate core networks. was made. During the that transition the mobile penetration rate was quite low, so many current subscribers of The third option is the common shared-network scheme, GSM were first time mobile subscribers and never expe- which can involve UMTS infrastructure. Operators build rienced a transition. Thus, little is known about the will- and operate a common 3G network, consisting of the ra- ingness of existing customers to invest in new terminals dio network and part of the core network. The base sta- to take advantage of a new technology. tions, RNCs, mobile switching centre/visitor location register (MSC/VLR) and serving GPRS support node (SGSN) A second difference between the GSM and UMTS rollout are all shared. is in the market structure. When GSM services were first offered in the Dutch market there was only one operator The final option is geographically split GSM and UMTS in the market. The UMTS rollout will occur in a market networks. Each operator builds and operates its own 3G with five licensed operators all of whom have several years network in different geographical areas. Each operator’s of experience operating in the Dutch market. The compe- subscribers can use the other networks via national roam- tition is expected to be high. ing agreements. When roaming in another network, the subscribers still have access to their operator’s services, whose name appears on the handset display [4]. 3.2.1 Business model There are several advantages which infrastructure shar- The greatest source of uncertainty in the 3G arena is which ing will bring to operators. Firstly, sharing reduces the business models will develop and succeed with mobile financial risks associating with the large initial capital data. The services based on mobile data are expected to investment required to build a UMTS network. Secondly, become a familiar part of many realms of people’s lives sharing could ensure that early 3G networks have the wid- from providing easy access corporate networks while trav- est possible geographical coverage, which will be essen- eling to helping organize their personal lives [2]. tial if 3G services are become popular. Thirdly, sharing Some of the key characteristics of mobile data services could help operators to meet regulatory commitments and will be positioning and personalization [5]. Positioning satisfy shareholders. Finally, sharing makes the planning or location dependent services, such driving directions, and designing of networks easily as there will be fewer parking or restaurant recommendations, will be possible cell sites. But infrastructure sharing could also add limi- as operators know the user’s locatiin by employing data tations to operators. It will decrease the independence and such as Cell ID, RTT-round trip time, and IP-DL (Idle flexibility of operators and increase the complexity of Period-DownLink). Services can also be personalized as compatibility if sharing partners with different equipment operators or service providers will be able to store and from different vendors. Table 2 depicts potential savings quickly access customer profiles. These features enable through infrastructure sharing. users to receive information that is relevant for them at the moment it is required. Thus, successful service pro-

28 viders will be required to provide timely information that is personalized and location relevant through a transaction process that is quick and easy to enact [6]. Figure 4. Simple 3G value Chain. Source: J.P. Morgan/ The challenge in creating business models for mobile data Arthur Anderson is how to bring various factors together. First one must integrate the featurrs of the UMTS technology such as its knowing the location of the user, the various connection speeds and the mobility they afford. Second, operators must be able to find valuable services and anticipate their prices for making revenue. Additionally, business models must fit with the strategy of the company and its target position in the market. Table 3. Revenue Streams Vary across the Wireless Data Industrial. Source: J.P. Morgan/Arthur Anderson (2001) 3.2.2 Revenue streams Existing operators with 2G networks who rollout UMTS Revenue streams for 3G services are expected to fall into networks will have to make significant changes in their 6 categories: person-to-person communication services range of skills to compete in the 3G environment. The (voice, email, SMS), advertising, commissions on trans- key assets of operators in this 3G business environment actions, business solutions, information services and surf- will be their micro payment billing infrastructure, a large ing, and entertainment (also referred to as multimedia). end user base, an established mobile brand, the users’ lo- A study [6] conducted by Nokia concludes that services cation information, established dealer channels and the will likely be offered in packages flexible combining fea- network infrastructure itself [6]. tures from four areas: information, communication, pro- If operators are able to offer services they will be able to ductivity, and entertainment. These dimensions reflect the enhance their position in the value chain. However ef- core uses of a variety of services and categorize them in forts to provide services will face hurdles such as lack of non-competing domains. They also reflect different tar- existing skills within the organization and a lack of capi- get markets. Business users may be interested in choos- tal to be able to develop or acquire such skills. Further- ing a package that specifies communication, information more, competition to play the role of service provider will and productivity enhancing services whereas the youth come from other market players [7]. market may be more interested in communication and entertainment. 3.3 Strategy The list of services included in Nokia’s study included email, ticket booking, maps, banking, weather, Internet As the market for mobile data evolves, it is a challenge browsing, traffic information, journey planning, shopping, for operators to develop strategies for a wide array of music downloads, video clip downloads, games and gam- business activities. In the immediate to near future im- bling. The list is arranged in order of the percentage of portant strategic issues respondents who indicated they are ‘very likely’ to pay include infrastructure sharing, access to networks and for such a service. The results range from 39% being very portals, and revenue generation. interested in email to only 2% who were interested in gambling. They reveal that market segments for 3G are starting to emerge, such as those interested primarily in games, 3.3.1 Infrastructure sharing music and video downloads, and others who are more Although the amount of cost reductions the unclear the interested in travel-oriented services. savings will help minimize the risk from scarce funding In 3G the revenue streams will also exist between differ- and slow revenue growth due to the huge investment and ent market players. Billing services or customer relation- uncertainty of the future. Infrastructure sharing also helps ship management provided to other application or con- meet regulatory commitments and satisfy shareholders. tent providers could be an important source of revenue. Expected benefits of infrastructure sharing also include Furthermore, operators may find location data or customer reduced hurdles in terms of site acquisition and pooling preferences data become products that can be sold to ap- of variety risks related to availability and compatibility plication providers. of equipment. Furthermore, infrastructure sharing could make the planning and design of networks easier as there will be fewer cell sites and fewer competing networks for 3.2.3 Value chain limited resources. The 3G value chain has the characters from both 2G value However, the immediate and longer term negative aspects chain and Internet value chain. A simple 3G value chain of infrastructure sharing should not be ignored. The stron- is showed in figure 4. Table 3 indicates revenue streams gest immediate negative aspect of infrastructure sharing for several market players. is the greater inter-dependency of one or more network partners. One negative aspect of this inter-dependency is

29 that it exposes partners to one another’s financial situa- power in GSM service begins its transition to UMTS with tion. If one partner is unable to meet cost sharing com- a closed access strategy they must carefully monitor their mitments, the position for the other partner, particularly market position [2]. given competitors’ successful sharing arrangements, will be weakened. But this is just one of many limitations that must be accounted for in the infrastructure sharing agree- 3.3.3 Revenue generation ment. Furthermore, issues of technical compatibility may Network services could be feasible sources for revenue complicate network rollout if partnerships are formed generation. First, focusing on network services is some- between operators with different vendors. thing within the realm of control of the operator. Second For long term business, infrastructure sharing is related for those operators lacking the financial resources to pur- to losses in competitive advantage. Operators not engag- sue external ventures, network services differentiation can ing in network sharing and with solid financial backing be developed in-house. Furthermore, if operators are to may be able to roll out networks quickly and use this com- pursue open access strategies, differentiated network ser- petitive advantage to increase market share. The size and vices could make them more attractive to potential con- ownership have implications on infrastructure sharing. tent and services partners. Focusing on network services Although the costs and benefits mentioned tend to be uni- may also provide the management structure necessary to form across operator groups, small operators may benefit closely manage the complexities of UMTS networks and more from infrastructure sharing if they are able to part- their effects on service levels [2]. ner with a larger operator. Larger operators could bring An especially important component of differentiation on more expansive embedded network resources such as cell network services will be managing the flow of informa- sites to the sharing arrangements. tion between the wide variety of players. In order to fa- Furthermore the larger firm’s influence with equipment cilitate the development of differentiation strategies and manufacturers who will be an integral part of the network to anticipate the various suppliers and customers of a par- rollout process may improve the overall position of smaller ticular service operator can refer to the value chain pre- players. A match between large and small partners could sented in section 3.2.3. This industry value chain can serve however complicate planning. Due to the relationship as a generic model of the processes required to offer a between capacity and cell size that results from the char- UMTS services. From it firms can develop more focused acteristic of the technology, the cell sizes needed by larger value chains for each separate service [2]. and smaller players could differ and sharing a network may result in non-optimal designs of investment decisions. 3.4 Cases The short and long term effects of such compromises require careful analysis [2]. In this section, information from two operators is collected. However, due to company confidential, only part of valuable information is publicly published. 3.3.2 Access

The second strategic component for operators is access 3.4.1 Sonera that is about how to allow the users to access the 3G services. Sonera is a leading Finnish communications operator, which provides communications services for consumer Open strategy is allowing the users to access the services and business customers. The 52.8% of the company are from any device and any other operators. The operator owned by Finnish State and the rest is publicly owned. has no any exclusive agreement with service provides. The close strategy is only allowing the users to access the It is estimated that by 2009 Sonera will spend 500 mil- services provided by the providers who has special agree- lion Euro in building its own UMTS network. The RAN ment with operator. is provided by both Ericssion and Nokia; the core network is provided only by Nokia. The benefit for the open strategy is the operators have wide sources to make revenue. However the negative The 3G services, planned by Sonera, are in 5 categories: things are that it requires extra skills and investment from messaging (email, text message and multimedia message), operators to deal with various service sources. This re- images (sending picture, text, graph, sound, video clips quirement could beyond the financial and organizational and real time video), content (delivering news and weather capabilities of the operators especially during the early report and updating every 15 minutes), locator (search- stage of UMTS rollout. ing for 18 different types of services anywhere in Finland according to current location) and game. In reverse, close strategy limits the revenue sources, but also reduces the skills and investment requirement. 3.4.2 ONI Way In addition to weighing the more general pros and cons of the open versus closed strategy, operators must consider ONI Way is a new entry operator, which does not have their own unique position in terms of ownership and size. GSM operation, in Portugal. The company is financed by For example, if a larger operator with significant market EDP, the national electricity company, in partnership with

30 Telenor (Norway), Grapes (Italy), Impresa (Portuguese In conclusion, the UMTS market is currently in a precari- media group), Brisa (national motoway company), ous situation. Given the sunk costs for license fees, the Jeronimo Martins (Portuguese retail distribution group) required further sunk costs for network infrastructure and and EFACEC (Portuguese manufacturer of electronic market development, network operators face considerable equipment). financial risk. A proper strategy is extremely important to reduce the existing risk. The UMTS infrastructure investment is about 1.6 billion GBP with the construction of 2750 base station. The equipment suppliers are NEC, Siemens and Nortel. REFERENCES [1] www.umts-forum.org 4 CONCLUSION [2] Johannes, M., Bauer, J., Rudi Westerveld, Carleen F. This report first outlined the background knowledge, Maitland. Advance Wireless Communications Infrastruc- which covers study background, UMTS background, the ture: Technical, economic and regulatory conditions of determinants of UMTS investment, the model of UMTS UMTS network deployment. Faculty of Technology, Policy market and the prediction of the UMTS development. A and Management, Delft University of Technology problem model of UMTS business from operators’ point [3] Lehman, B. (2001). Wireless Services: Industry Over- of view was also introduced. The model reflects the real view. London, Lehman Brother Equity Research: 9. situation of UMTS investment, which is expensive and uncertain for the future. The following report records the [4] Torbjorn, P. Sharing Agreements: Operators Must Look result of the study that is based on this problem model. Before They Leap. Ericsson. In the result of the study, the factors that affect the UMTS [5] Borrel, J. (2001). Gearing Up for 3G upside 13(3): investment were discussed. These factors include cost, 84-89. revenue and operator strategy. The cost contains license [6] Nokia (2000). Make Money with 3G Services. Espoo, cost and infrastructure cost and both of them were ana- Finland, Nokia. lyzed in detail. Also the cost saving through infrastructure sharing was emphasized. To revenue, the possible [7] Dundee (2000). The Services and Technology of the 3G revenue stream and value chain were indicated. After Wireless Internet: Costs and Benefits. Toronto, Dundee that, two operator strategies about infrastructure sharing Securities Corporation: 60. and access to services were analyzed as well.

31 CDPD Service

FENG YUE LIU SHUPING DEPARTMENT OF ELECTRONICS AND COMMUNICATION DEPARTMENT OF ELECTRONICS A ND COMMUNICATION FIN-02150 HUT, ESPOO, FINLAND. FIN-02150 HUT, ESPOO, FINLAND. [email protected] [email protected]

ABSTRACT Speed Wireless data service is more and more demanded now. CDPD’s raw transmission rate can be 19.2 kbps, which is CDPD (Cellular Digital Packet Data) service is one of faster than any other available data service ten years ago them, which is very popular in North America, Latin, South when it was born in 1994 ( refer Appendix 1). America and some Asian Pacific countries. In wide sense, Open Specification it can be viewed as a kind of service, technology, standard and network. In this sense, CDPD’s success is also CDPD is based on one of the most widely accepted net- because of the factors mentioned above. Nowadays, many working protocols TCP/IP, it provides very compatible public and professional services are based on CDPD. It is service with other existing IP-based applications in a wire- specified as a packet-switched technology to support wireless environment. This also provides the possibilities of less access to the Internet and other public packet-switched CDPD mobile subscribers’ connection to the Internet via networks. Because of its’ open characteristics that adheres CDPD by economic, small latency, reliable, easier means. to the layered structure of the Open Systems Interconnec- Reliability and Security tion (OSI) model, it has the ability to be extended. CDPD implements the FEC mechanisms reducing the transmission errors over the Air Interface. Authentication 1 INTRODUCTION and encryption to all packets transmissions are also implemented over the Air Interface. Currently, mobile data communication services have de- Availability veloped significantly. Generally speaking, mobile data communication services are realized by several cellular Since CDPD utilizes existing cellular networks, carriers technologies. They are Circuit-switched cellular technol- can deploy CDPD services in a short period of time to ogy, Packet-switched cellular technology and Cellular meet customers’ specified coverage requirements. The Digital Packet Data technology. Actually, in some other general CDPD platform allows the easy integration with essays, both Cellular Digital Packet Data technology and existing applications, and a significant capability to pro- Packet-switched cellular technology belong to the same vide wireless Internet accessibility. This is partially be- category of Packet Switched Technology, because their cause the CDPD is a compatible and open standard. principles behind are same. The CDPD service we are Cost Effectiveness talking about is based on this. CDPD uses idle resource of AMPS and Packet Switched CDPD is a wide-area network enabling the transmission technology. So it is cheaper. of packet data across the cellular networks. The network is designed to be an extension of existing TCP/IP data communication network via wireless access from subscribers. It is built on the top of AMPS (Advanced Phone Sys- 2 CDPD HISTORY tem), where CDPD obtains an idle 30 KHz channel and CDPD service is based on the 800MHZ AMPS, which transmits data frames at 19.2kbp [1][2]. CDPD shares the was born in 1980’s. It says, without AMPS, it goes with- same radio frequency channel with AMPS cellular net- out saying what CDPD is. work on the Air Interface. However, its’ own infrastructure for transmission over AMPS is required, which can In 1991, CDPD patents were filed by 3 IBMers (Miller, deal with data transmission and AMPS speech streams Moore, Pate, IBM Boca Raton). separately. Obviously, we can get the conclusion, the In 1992, several North American telecommunication gi- CDPD was totally tailored for American market. ants, AT&T, MGcaw, BAM, SW Bell, North Telecom, Today’ many technologies and service are not in the shade, Ameritech invested $1000,000 each as a fund to make a compared with CDPD. What benefits can it bring for the research of offering packet data service. It was so fast customers? that, in 1993, the first version of standard of CDPD was released. In 1994, in America, the first CDPD experimental network run. In the second season of the same year,

32 several telecommunication service giants, jointly an- say CDPD provides a packet-switched technology. Un- nounced they were founding on a plan of CDPD service like the circuit switched service, CDPD service reuses which is totally customized in United States and also cov- the unused voice channel by hopping technology. ered the whole North America as soon as possible. Fol- Channel Hopping lowing that, several American countries have selected the CDPD to offer packet data service. Over the Air Interface, CDPD uses the idle capacity by the configuration of “sniff and hop”. Two types of hop- A milestone of CDPD development is the establishment ping mechanisms come [1]. of CDPD forum. The CDPD Forum was founded in April 1994 to provide a more structured way for companies in- Planned channels hopping: It occurs at a given time that volved in CDPD to work together. It was responsible for the CDPD specifies. CDPD network management func- the CDPD standardization, specification release, the lat- tion configures Base Station to use a specific channel only est technologies and applications in this field and market for a given period and then it must capture another radio analysis. This forum did contribute a lot for the develop- channel in a round-robin fashion. It is also the Base ment of CDPD, especially the compatibility solution. Station’s task to direct the user’s modem to a new chan- According to the statistical records from the CDPD fo- nel to be used for CDPD activity. rum, there are about 50 metropolitan areas in America Forced channels hopping: When a cellular active voice provided this service by the top 39 biggest operators in activity is detected on a channel that it is currently used North America. Totally about 21 manufactures have signed for CDPD service, the CDPD Base Station will change it alliance of CDPD with agreements of promising to pro- to another frequency channel where there is no CDPD vide compatible equipment. speech service is using. This is helpful to avoid the inter- With the increasing growth of the wireless data business, ference between the CDPD data service and AMPS speech a more international organization, Wireless Data Forum service. However, even the hopping is in progress, like was born. Wireless Data Forum takes place of CDPD the other modern mobile services, it is transparent to the Forum, which embraces all wireless data technologies subscribers. Otherwise, the CDPD service will be unbear- including CDPD related matters. able for mobile subscribers. TCP/IP Based 3 CDPD FEATURES CDPD service is pretty clever in this sense that it is TCP/ IP protocol based. Both in market term and in technical Number Of Users Allowed term, this provides possibilities of success for CDPD. The CDPD system data link has a maximum capacity it There is no needs of extra effort of protocol translation can support. The Air Interface of CDPD is the main lim- which costs extra hardware, cost, slows speed and reduces iting factor of the capacity of the network because the reliability, when connecting to the Internet and other TCP/ 19.2 kbps maximum throughput of CDPD limits the num- IP network. ber of frames sent over the channel. However, CDPD ra- CDPD Transmission Rate dio data link can support dynamic number of users. CDPD raw transmission rate is 19.2kbps. However, if error FDMA Based CDPD control overhead is considered, the actual throughput is CDPD is on the top of the AMPS network which is FDMA only 12kbps on a so called clean and lightly loaded chan- characterized. Inheriting from AMPS, CDPD is FDMA nel. In order to increase the throughput, it also compresses based, even though it is digital and packet data service. protocol overhead. CDPD does not need the TCP/IP con- Unlike today’s other digital TDMA system, CDPD uses version for the Internet connection. Further, it even com- whole frequency channel during the time it is allowed. presses the IP protocol overhead, increasing the through- However, in TDMA system, several channels share one put to make the frequency bandwidth efficiency higher. frequency channel at non overlapping time. Compared Encryption And security with TDMA system service, CDPD, this kind of FDMA based service, has lower frequency efficiency. In order to prevent unauthenticated subscribers’ access to the CDPD network, the subscribers´ identity will be veri- CDPD Packet-switched Transmission fied whenever they connect to CDPD network. This veri- CDPD operates over AMPS analogy cellular telephone fication process is actually the verification of MES iden- network, whose voice call service employs the radio fre- tify [1][2]. During the time of dialogue between the MES quency of 824-894 MHz. CDPD is fully digital, which and CDPD network, the responsible MDIS requests elec- uses the GMSK modulation to encode data. tronic key exchange. MES provides a triple set of NEI, ASN and ARN to the network. The MES Home MDIS Actually the main theory behind of additional revenue will verify this triple set to accept or refuse the access of earned is that CDPD uses the short blank spaces between the requesting MES. Each time a subscriber turns on his/ the AMPS voice calls, which results in improving the re- her MES, MES will send a request packet to the Base source efficiency. CDPD subscribers send the packet ori- Station to the MDBS which forwards it to MES’ Home ented data on radio channels within a sector, when they MDIS. Then the Home MDIS will forward it to the AS are not currently used by the AMPS voice call. So we can server which is responsible for MES identify verification.

33 About the CDPD service encryption, all data transferred tween X.25 and IP, and X.25 packet and IP packet. CDPD between CDPD modem and MDIS must be encrypted by can also internetwork with PSTN. RSA algorithms. However, the data beyond the MDIS is not encrypted, they are like normal Internet traffic unless the end users require further encryption. 4 CDPD NETWORK COMPONENTS Access Control And Congestion CDPD operates as an overlay on top of the AMPS net- CDPD uses means of DSMA-CD (Digital Sense Multiple work. Thus for implementing the CDPD, only 4 main sub- Access, Collision Detection) as its’ access control and systems in the CDPD network, ES, MDBS, MDIS, and congestion mechanism. It is like the normal Ethernet IS are needed. The CDPD network topologi is presented CSMA-CD (carriers sense multiple access, collision de- in Figure 1. tect). Before a device sends its’ data, it will sense to see whether the channel is busy. If not busy, the device will send the data. However, If another device acts at approximately same time, a data collision will occur. So the sending device needs to wait a random time to try again. The CDPD also defines MAC (Media Access Control) protocol which manages the function of DSMA-CD. In order to avoid that a device occupies a channel for too long time, the MDBS can set a maximum time for a device accessing to a channel, in general for single CDPD transmission, it can only hold a radio channel for 1 second. CDPD network allows a maximum of 64 blocks or 385 Figure 1 bits each in a single transmission. But in real world of MDBS (Mobile Data Base Station transmission, the actual limit is set by a parameters controlled by MDBS. Mobile Data Base Station. It is responsible for the Air Interface control, radio frequency management and chan- Cell Handover And Roaming nel hopping. It acts as the Base Station in GSM network. Like other modern data service. CDPD also supports the A Interface - Air Interface cell roaming. It has a completely elaborate protocol to support the roaming management. When a MES’s receiv- The interface, between the Mobile End Station and cell ing signal strength is lower than a specified threshold, or site equipment MDBS, provides the function of wireless Block Error Rate is higher than bearable threshold, even diversity in order to improve the radio reception. the MDBS is unsynchronized, MES will initiate cell MDIS (Mobile Date Intermediate System) handover request. Cell handover and roaming mainly depends on the protocol of CLNP between different MDIS It uses the powerful computers and internetworking tech- [1]. nology to deal with requests from the CDPD network also connected with other MDIS. It includes two functional Operators have roaming agreement with each other. The parts of Packet Server and Admin Server. The Admin CDPD subscriber’s roaming area is very large. Mobility Server can select the Centralized Control or Decentral- Management Service maintains the current location in- ized Control as its’ implemented means. formation of the CDPD subscriber. For the subscriber, there is a seamless data connection for them as they are MDIS controls the several MDBS using high speed dedi- roaming in the serving areas. However, all procedure is cated channel of 64kbps, the geographical factor of the transparent to the CDPD subscribers. each cell will be used to decide which MDIS it belongs to. All these are quite like the GSM system, except CDPD Numbering Mechanism does not have some elements which can match the BSC. CDPD network is TCP/IP based network. It provides stan- In the GSM system, the BSC will care the radio exchange, dardized TCP/IP interface to external TCP/IP network. setting up the radio channels for the traffic and for signal- Then IP address comes. Because of lacking IP address, ing to MSC. In CDPD, it integrates this function to the most CDPD network operators use the reserved IP ad- MDIS whose function can partially match the MSC in the dress. However, it needs gateway to make the IP address GSM system. It results in bringing more load to the MDIS. translation when connecting to external Internet. Prob- MDBS holds the role between the M-ES and MDIS. MDIS ably the gateway will bring bottleneck of traffic between is the only network element which knows the M-ES mo- these two networks. bility in the system. Internetworking With Other Network. IS (Intermediate System) CDPD network also provides possibilities for subscrib- CDPD is an IP based system. CDPD backbone network ers to communicate with the X.25 network with the help is construed by the IS. Actually they are the routers. It of router which functions as a address translation between provides the connectionless data grams service. Accord- X.121 address and IP address, protocol translation being to the IP address, it can provide the function of ad-

34 dressing. Mainly the IS uses the protocol specified by network layer. Network layer functions provided within the CDPD network is used to communicate with each other. When the data grams are transmitted, there must be a path determined, if interconnected elements called IS. So we can easily recognize that a IS is a router, to a certain extent[2][3]. CDPD ES Not in a strict sense, CDPD subscribers use two kinds of ES (End System), Mobile End System (M-ES) and Fixed End System (F-ES) to access CDPD network [2][3]. In order to manage easily, the CDPD system separates Figure.2 them. M-ES is somehow like a wireless portable PC which can travel from one cell to another and communicate with As the figure 2 shown above, a small chain shop does not CDPD network via MDBS[2][3]. need build a LAN. CDPD can provide a nice solution. To Simply speaking the F-ES is whatever network environ- connect the existing cash register with a CDPD Modem ment/resource the subscriber’s equipment is attempting or use the wireless POS computer. It is more convenient to make available to mobile access via CDPD. External than the conventional solution of connecting to the bank. F-ES are administered by CDPD subscribers. Internal F- ES are administered by the CDPD operators such as AT&T. The F-ES could also be the directory services da- 6 DISADVANTAGES OF CDPD TODAY tabase. It is important to realize that the F-ES is not re- Speed is slow. Today’s GPRS speed is 171.2kbps quired to be aware of mobility issues in any way whatso- (21.4kbps per channel) and CDPD is about 19.2 kbps (re- ever. As such, an F-ES should be able to connect to a fer Appendix 2). CDPD network with absolutely no modifications. Capacity of CDPD is relatively smaller. Frequency efficiency is lower compared with other Euro- 5 APPLICATIONS OF CDPD pean system, such as GPRS, GSM and HSCSD. Each Transportation carrier frequency can take 8 slots at most. However, CDPD can not. Some companies, such as lorry transportation companies, express post companies and taxi companies, can estab- CDPD can only offer the connectionless service. On mar- lish communication between their dispatchers and the ket, this is an obvious disadvantage of CDPD [5]. moving cars, so as to carry out real-time control and real- GPRS has the QOS level based on the levels of priority, time management of their cars with mobile packet ex- delay, reliability, input/output. However, CDPD does not change network [4]. have the QOS level. CDPD can not provide as more Public security choices to subscribers as GPRS can [5]. With the mobile data communication, the public security personnel can contact remote databases at any time to get 7 CONCLUSION the latest information concerned about the missing cars, the criminals and finger marks. CDPD is a transition solution until the CDMA2000 is really realized. Here the authors emphasize the CDMA Commerce 2000 system, because both of them are originated the With the progress of commerce electronicization and the North American mobile communication standard. So still development of the gold card project, the credit card at- CDPD will be active for quite certain while. It faces the testation system and the automatic drawing machine will big competition from GPRS now. Also the new comer of be developed greatly. The mobile terminals of this kind Wi-Fi makes the competition much harder. CDPD prob- of system are easy to be erected and convenient to be ably will lost part of the telecommunication pie of private used, so they can be used easily in sports meetings and users. However, its’ life will be longer in the professional commodity fairs. Also the wireless ATM realized over networks communication. We can not negate that CDPD CDPD is coming. The figure below is an example of wire- is a good system. However, it is a little bit of out of date less casher via CDPD. now. Even in its’ birth place, United States, more and more its’ previously loyal subscribers turn to other data service. But within 3 to 5 years, it will not disappear from our eyesight.

35 8 REFERENCE [1] http://www.sierrawireless.com/pub/doc/2130006.pdf, CDPD Primer. [2] http://www.soloist.com/CDPD.htm [3] Page 454- 456, Wireless Communications Principles and Practices. Theodore S. Rappaport. 1996 by Prentice- Hall. [4] http://www.cdpd.org.cn/index1/FIle/GPRS_point.htm [5] http://www.cdpd.org.cn/index1/gps_simple.htm

Appendix 1: Comparison between the CDPD service and first and second generation mobile service.

Technology Reliability Coverage Mean data Equipment and Security Mode Analogous Voice or data rate (kbps) communication costs or digital support

First generation Average Good 9.6 Average Poor Duplex Analogue Both cellular

Second generation Excellent Good 9.6 Average ~ good Excellent Duplex Digital Both cellular

CDPD Excellent Average 19.2 Excellent Excellent Duplex Digital Data

Appendix 2: Comparison between the CDPD service and GPRS service.

GPRS CDPD System Type GSM Based AMPS Based Transfer Speed 115kbit/s (14.4kbit/s/channel) 19.2kbit s Switching mode Packet Packet Bandwidth 200Khz 30Khz Protocol TCP/IP TCP/IP Channel Shared Shared Capacity The number of data service user is 15000 users can login in simultaneously restricted by voice traffic. for each PS Delay Short Short Roaming Capabilities Supported Supported Charging Rate Charged by packets Charged by packets

36 5 GHz WLAN

TAPIO SOKURA TELECOMMUNICATIONS SOFTWARE AND MULTIMEDIA LABORATORY HELSINKI UNIVERSITY OF TECHNOLOGY FIN-02015 HUT, ESPOO, FINLAND [email protected]

ABSTRACT 2 BEFORE 5 GHZ This is a short introduction to the evolution of Wireless In this section we are going to take a look on current and Local Area Networks. Special consideration and compari- historic wireless LAN technologies that operate on fre- son is done on the 5 GHz WLAN standards, IEEE 802.11a/ quencies below 5 GHz. h and ETSI HiperLAN/2. 2.1 The Proprietary Ones 1 INTRODUCTION WLAN systems that resemble the currently standardized WLANs (access points connected to the wired network, The continuous move towards using higher and higher portable clients etc) were available from some manufac- frequencies and bandwidths–no matter what the technol- turers in the middle of 1990s, each having their own pro- ogy–is usually driven by the need for higher capacity, prietary technologies. The systems were rare, quite ex- which in the digital world is faster transfer speed. One pensive and low speed. For example Lucent provided a example of the growing need for capacity can be seen in system that operated at around 900 MHz and offered a table 1 that lists Finnish mobile phone standards from transfer speed of 2 Mb/s (megabits per second). 1970s ARP to the UMTS of 21st century. Often having a common standard is an important part of making a technology success. This is very true also in the Standard Frequency Bandwidth WLAN world, where the IEEE 802.11 standard defined in 1997 pawed the way for having different manufactur- ARP 150 4 ers’ products interwork. NMT 450 10 NMT 900 20 2.2 IEEE 802.11b, aka Wi-Fi GSM 900 50 IEEE 802.11 is the first of the IEEE wireless LAN stan- GSM 1800 150 dards and the radio part uses the license-free 2.4 GHz ISM-band available worldwide. It provides a maximum UMTS 2000 155 transfer speed of 2 Mb/s and supports up to 14 channels. Table 1: Frequency and bandwidth of mobile phone Not all of the channels are available everywhere in the technologies (in MHz) [UMTSFIN]. world, but most countries support a common subset that enables worldwide compatibility, in theory at least. Having a common stan- The same desire to achieve higher capacity, both in speed dard didn’t automatically guarantee seamless interworking and in simultaneous users, has driven the Wireless Local between different manufacturers’ products. But it was a Area Network (WLAN) manufacturers to higher frequen- step to the right direction. cies, where more bandwidth is available. 802.11b upgraded the older standard to a speed of 11 Mb/ This report focuses on the latest WLAN standards that s in 1999, using DSSS (Direct Sequence Spread Spec- operate on the 5 GHz band, namely IEEE 802.11a and trum) modulation. There are 14 channels in the b-stan- ETSI HiperLAN/2 (European Telecommunications Stan- dard as well, but because neighboring channels overlap dards Institute). Also some previous and other rivaling each other, only 3—5 can be efficiently used in the same technologies are presented, for example IEEE 802.11b area to avoid interference. The b-standard is where and g that operate on the 2.4 GHz ISM-band. Differences WLANs really took off and it is the most popular WLAN between these technologies are discussed and how they technology today. The formation of the Wireless Ethernet might affect future deployment of WLANs. Compatibility Alliance (WECA) in 1999 also played an important role in the success of WLANs. WECA — nowadays known as Wi-Fi Alliance — is a non-profit organi-

37 zation that certifies 802.11b compatible products, so that 802.11g 2.4 54 3 OFDM no customers can be reasonably sure that the product they 802.11a,h 5 54 19 OFDM no are buying indeed works with other manufacturers’ products. There was clearly a need for the certification be- HiperLAN/1 5 23 5 GMSK some cause early 802.11(b) products of a certain manufacturer HiperLAN/2 5 54 19 OFDM yes were often partly or fully incompatible with the products of a competitor despite both of them implementing the HiperLink 17 150+ ? ? yes same standard.

The current install base of 802.11b compatible devices is Table 2: Some WLAN standards’ features. The amount huge and it is still the best selling WLAN technology. It of channels available varies between countries and will therefore probably be in use for several years to come. IEEE 802.11b supports only 3—5 non-overlapping Compatibility with previous standards is an important channels. point to keep in mind when guessing the future WLAN success stories. 3.1 Why 5 GHz? 2.3 IEEE 802.11g The switch to 5 GHz is not an easy one to do, neither A new standard for 2.4 GHz WLANs is currently in the technically nor financially. The area that one 5 GHz final draft stages in the IEEE standards process, named WLAN access point can cover is significantly smaller than 802.11g. Even though there are already devices claiming the area a single 2.4 GHz AP can cover. This means that “802.11g compatibility” in the market, the standard has for equivalent coverage the 5 GHz network needs more not yet been finalized and it is subject to further changes. access points, which translates into higher costs. Because The final standard is expected to be ratified in June 2003 the frequency band changes, WLAN users also need new [80211g]. equipment to take advantage of the new access points. Dual-band access points are one way of expanding the What makes 802.11g interesting, is that it is downwards coverage for both old and new standard users, but they compatible with IEEE 802.11b and provides a maximum are usually more expensive than single-band APs. transfer speed of 54 Mb/s between two 802.11g nodes. The modulation method and the maximum speed are the An access point is the device that links the wireless LAN same as with the currently specified 5 GHz WLAN stan- to the wired local area area network (LAN). Basically an dards, IEEE 802.11a and ETSI HiperLAN/2. Because the access point in WLAN is like a base station in a mobile frequency band of the g-standard is lower, there are only phone network. three channels available. On the other hand the expected Regulations governing the use of 5 GHz band differ more coverage area of a single access point is considerably between countries than they differ on 2.4 GHz ISM-band. larger than on 5 GHz. This has lead to a situation where one technology can be 802.11g will probably be of special interest to those who used on one continent, but not on another. Typical differ- need to expand WLAN coverage area or upgrade network ences include different band edges, different maximum speed, and have a wide installation base of older 802.11b power levels and possibly some special requirements, such equipment. With the g-standard equipment it’s possible as automatic channel selection. Table 3 lists common un- to do both of these in the same time, although overall licensed WLAN frequencies in the USA, Japan and Eu- network capacity may become a problem due to having rope. only three “g-channels” available. Frequency Europe USA Japan (GHz) 3 AT 5 GHZ 0.902—0.928 not allowed 1W not allowed Problems such as interference from other ISM-band devices and capacity/speed shortage with the WLAN stan- 2.4—2.4835 100 mW 100 mW 100 mW dards that operate on 2.4 GHz is slowly leading to the 5.15—5.25 200 mW 50 mW indoors only adoption of 5 GHz band WLAN standards. Table 2 lists indoors only indoors only some attributes of the discussed WLAN standards. This section will mostly deal with IEEE 802.11a and ETSI 5.25—5.35 200 mW 250 mW not allowed HiperLAN/2 standards. indoors only 5.47—5.725 1 W not allow. not allowed

Standard Freq Speed Channels Modulation QoS 5.725—5.825 not allowed 1 W not allowed GHz Mb/s (max)

802.11 2.4 2 14 PSK no Table 3: WLAN frequencies including rough power and 802.11b 2.4 11 14 CCK no usage limitations in the USA, Japan and Europe (Finland).

38 Despite the above, using 5 GHz for WLANs does have But this is not enough as the regulations in most Euro- advantages over 2.4 GHz. One of the biggest benefits is pean countries also require at least DFS (Dynamic Fre- the considerably wider bandwidth available. For example quency Selection) and TPC (Transmit Power Control), in Europe the WLAN-accepted 2.4 GHz ISM-band is 83.5 which aren’t part of the 802.11a standard. The reason for MHz wide, but on 5 GHz there’s 455 MHz allocated for these requirements is that there are satellite links in use WLAN use. Another advantage is that the spectrum used on the same frequency band in Europe. To keep interfer- for WLANs around 5 GHz is largely not in use at the ence to these links as low as possible, these two require- moment, so interference from eg. Bluetooth equipment ments have been specified. or microwave owens is not a problem. Also the OFDM When an access point uses DFS, it selects the best operat- modulation used by both 802.11a and HiperLAN/2 is more ing channel automatically depending on other traffic and robust than the modulation methods used by older 2.4 interference levels. A node that uses TPC adjusts the trans- GHz standards. mit power level depending on how well it’s transmissions are heard by its communication partners. The idea is to 3.2 IEEE 802.11a use the minimum power level that is sufficient to sustain the communication, to decrease interference to other de- IEEE 802.11a is basically a remake of 802.11b’s physi- vices on or near the same frequency. Even if a certain cal layer for the 5 GHz band. The a-standard was devel- 802.11a product fulfills these requirements, it currently oped in parallel to 802.11b and they were ratified at the has to be individually approved in each European coun- same time in 1999. The only significant differences are in try before it can be legally used. the operating frequencies, transmission speeds and in the modulation methods used. This also means that the security vulnerabilities of 802.11b’s WEP-encryption are also 3.3 ETSI HiperLAN/2 present in the a-standard. The European Telecommunications Standards Institute is MAC (Media Access Control) is done by the use of — just like the name suggests— the European standard- CSMA/CA. This basically means that before transmitting ization organization for telecommunications. This is dif- a data packet the node broadcasts a packet that warns other ferent from IEEE, whose focus is more on the users that there’s going to be traffic from me. After this datacommunications side. the node sends the actual data packet. If a collision hap- So what happens when a telecommunications organiza- pens, then it happens and the node has to try again after a tion decides to write a WLAN standard? You get a stan- certain timeout expires. dard that is biased towards the telecom-world, and that’s exactly what HiperLAN/2 is. Someone might think that 3.2.1 Physical layer this is a burden for HiperLAN/2, but in fact this turns out to be mostly an asset. HiperLAN/2 standard was published Instead of using DSSS, 802.11a uses OFDM as it’s modu- in spring 2000 and it aims to be a general solution for lation method. OFDM splits the 20 MHz wide channel high speed mobile data transfer, not just another wireless into 52 distinct subchannels. Splitting the single fast stream LAN technology. of bits into several slower streams reduces problems caused by multipath propagation, where the same signal HiperLAN/2’s physical layer is just about the same as is arrives at the destination via two or more different paths in 802.11a: 54 Mb/s OFDM. The exact frequencies used and thus on differ a bit, but the rest of the physical layer is the same. different times. Retrieving the original data from this combination of mixed time signals is considerably harder and 3.3.1 Differences to 802.11a sometimes impossible. The actual differences are found on top of the physical An 802.11a compatible device has to support data rates layer. The most significant differences between 802.11a of 6, 12 and 24 Mb/s, it may optionally also support data and HiperLAN/2, in no particular order, are: rates of 9, 18, 36, 48 and 54 Mb/s. Because the maximum - QoS (Quality-of-Service) support transfer speed of 54 Mb/s is an important sales point, most normal WLAN equipment will probably support all of - advanced authentication & encryption support the abovementioned speeds. - dynamic frequency selection & transmit power control - connection-orientation 3.2.2 Clashes with European regulations - better compatibility with ATM, Firewire, UMTS... There are 12 distinct channels in use for 802.11a in the USA. Eight of these fit inside the European bandplan and can be used in Finland. In addition to these common bands, The MAC layer in HiperLAN/2 is significantly different there are other frequencies set aside for WLANs in Eu- from the traditional Ethernet world, be that wired or wire- rope that are used for something else in the USA and vice less. Instead it resembles ATM in the sense that all data versa. transferred over a HiperLAN/2 link is split into 48 byte

39 cells. It resembles GSM in the sense that there are spe- I doubt that there will ever be a real union or merge of cific timeslots for specific kinds of messages. For example, 802.11a/h and HiperLAN/2 standards, I think that’s not if a client (MT, Mobile Terminal in telecomms terms) the point. The idea is to standardize on the frequencies needs to establish a connection with the access point, the and physical layer so that remaining compatibility require- client contacts the AP using a random access channel ments can be met in soft/firmware, if needed. HiperLAN/ (timeslot) reserved for this very purpose. When the con- 2 is as strongly based on ATM and other telecommunica- nection is accepted and formed, the client gets its own tion technologies as 802.11a is on wired Ethernet tech- regular timeslot that it can use to send data to the access nologies. point. Because the data transfer is controlled by the AP and every node gets it’s own timeslot, there are virtually no collisions. This method of allocating frequency space Standard Description / Objective to multiple users is known as Time Division Multiple 802.11 The first IEEE WLAN standard, Access (TDMA). 2 Mb/s on 2.4 GHz DSSS/FHSS/IR The use of TDMA and connection-orientedness of 802.11a 54 Mb/s on 5 GHz OFDM HiperLAN/2 enable QoS support for applications that need a steady data flow with a fixed maximum delay, such 802.11b 11 Mb/s on 2.4 GHz DSSS as voice- or video-conferencing. Also the traditional 802.11d To get 802.11 permitted in countries that Ethernet “best-effort” type of delivery is available, if don’t accept it at the moment needed. The access point keeps score of how many clients it has and what are their QoS requirements to be able 802.11e QoS for WLANs to decide whether a new connection can be allowed or if 802.11f Roaming between access points it should be denied to keep the QoS level of existing clients acceptable. 802.11g 54 Mb/s on 2.4 GHz OFDM HiperLAN/2 also supports one- and two-way authentica- 802.11h The addition of DFS and TPC to 802.11a tion via PKI and strong cryptography via 3DES encryp- to legalize it in Europe tion. Basically this means that it should be easier to build 802.11i To improve WLAN security with TKIP, secure wireless networks with HiperLAN/2 than it is with 802.1x and EAP the IEEE standards. But because there are practically no actual HiperLAN/2 products on the market, it’s impossible to say how it would work out in practice. Table 4: A summary of related IEEE past and present Despite all of these enhancements compared to 802.11a, 802.11 Task Groups. HiperLAN/2 products aren’t expected to be more expensive than 802.11a products. At this stage this just specu- lation because there seems to be no ready HiperLAN/2 4 WHICH ONE TO CHOOSE? products available on the market. At least I didn’t find any. This is a bit surprising as there are many big names For attracting the largest WLAN user base 802.11b is the standing behind HiperLAN/2, for example Nokia and clear leader at the moment, there’s no question about it. If Intersil. we think of speed and capacity, the HiperLAN/2 and 802.11a are fastest and provide the most capacity compared to other WLAN technologies. 3.4 A Global Standard? In absolute speed HiperLAN/2 is somewhat faster than Because IEEE 802.11a is facing regulatory problems on 802.11a because HiperLAN/2 uses fixed size cells that it’s world tour (read: Europe), IEEE founded a new 802.11 always have the same amount of overhead. 802.11a/b/g/h Task Group, designated with the letter h. 802.11h stan- use variable length packets, which means that they slow dard is supposed to eventually be a drop-in replacement down when transferring relatively small packets. On the for the a-standard and they will most likely interwork fine other hand, small packets tend to be sent when there’s not with each other. that much traffic on anyways, so this might not be an issue. It could become in issue in applications that require The h-standard incorporates the transmit power control real-time performance from the network, such as voice and dynamic frequency selection features of HiperLAN/ conferencing. Of course there’s the 802.11e Task Group 2. Other small changes are possible, but none are expected. working on QoS for the WLAN technologies, but it’s not The h-standard is estimated to be finished by the end of ready yet. 2003. The Japanese also have a system of their own for high-speed wireless networking called MMAC (Multime- 802.11a/b don’t offer much security with their compro- dia Mobile Access Communication). The current situa- mised WEP-encryption and 802.11i that would enhance tion seems to be that the physical layers of all of these the security of existing standards (802.11a/h, HiperLAN/2, MMAC) are in prin- 802.11 standards is also still work in progress. At the ciple compatible: 5 GHz OFDM. moment it would probably be best to use IPSec or some other higher layer all-encrypting VPN with mutual au-

40 thentication for sensitive wireless LANs. If security or CCK Complementary Code Keying QoS needs to be embedded in the product, then the only CN (UMTS) Core Network current choice is HiperLAN/2. But then again, where are CSMA Carrier Sense Multiple Access the products? The technology behind HiperLAN/2 seems to be solid in many ways, but it’s of no use without work- DFS Dynamic Frequency Selection ing products. DSSS Direct Sequence Spread Spectrum 802.11g may turn out to be a tough competitor for 802.11a EAP Extensible Authentication Protocol because of the built-in compatibility with 802.11b. Also ETSI European Telecommunications Standards Institute the price of g-standard products can easily be lower than GMSK Gaussian Minimum Shift Keying of those 802.11a products that also have a b-standard radio in them. We’ll probably see 802.11a/g combination GSM Global System for Mobile Communications access points and cards, they would have the best com- IEEE Institute of Electrical and Electronics Engineers patibility with existing and new WLAN equipment. A weak ISM Industrial, Scientific, Medical point of 802.11g compared to the 5 GHz technologies is LAN Local Area Network its lower capacity; only 3 channels are available vs. at least 8 and over on 5 GHz. MAC Media Access Control Mb/s megabits per second MMAC Multimedia Mobile Access Communication 5 CONCLUSIONS MT Mobile Terminal If and when 802.11h standard is agreed upon this year, 5 NMT Nordic Mobile Telephone system GHz WLAN will finally become reality in Europe. But OFDM Orthogonal Frequency Division Multiplexing the change from 802.11b won’t happen overnight or overyear. Older 802.11b equipment will stay around for a PSK Phase Shift Keying long time, after all its speed is more than enough for the QoS Quality-of-Service typical laptop WLAN user, not to mention a typical PDA TDMA Time Division Multiple Access WLAN user. TKIP Temporal Key Integrity Protocol The support and choices of major WLAN vendors will TPC Transmit Power Control naturally play a central role in deciding which protocol UMTS Universal Mobile Telecommunications System will be the dominant one. And those fingers are definitely pointing to the 802.11 series of standards. Whether it will WLAN Wireless LAN be 802.11g or 802.11h in the long run is hard to say. Maybe the h-standard will stay as an indoor standard due to it’s shorter range and the g-standard slowly conquers the outdoors from 802.11b. HiperLAN/2 seems to be a solid tech- REFERENCES nology, but nothing is happening. It looks like HiperLAN/ [BRAN] ETSI Broadband Radio Access Networks, 2 is in the end going to be just another standard born dead http://portal.etsi.org/bran or killed by it’s ex-supporters. HiperLAN/2 might gain some ground on telecommunications type applications [HiperLAN2] HiperLAN/2 Global Forum, http://www.hiperlan2.com (real-time audio and/or video etc) in Europe. But it won’t beat the IEEE standards in “regular” LAN access, espe- [UMTSFIN] Finnish Communications Regulatory Authority, cially in the USA. UMTS-taajuudet ja toimiluvat Suomessa, 30.11.1998. Online, referred to 22.3.2003, http://www.ficora.fi/suomi/document/ finalumts.pdf GLOSSARY [WECA] Wi-Fi Alliance. http://www.weca.net 3GPP 3rd Generation Partnership Project (UMTS) [WGloss] Wireless Glossary. AP Access Point http://www.enterprise-zone.com/wireless/Door/11259 ARP Autoradiopuhelin (“car radio phone”) [WLANs] UNINETT: Wireless Networks, http://www.uninett.no/wlan ATM Asynchronous Transfer Mode band frequency range, for example 2.14—2.27 GHz [80211g] IEEE, IEEE 802.11g Gains Working Group Approval, BRAN Broadband Radio Access Networks 14.2.2003. http://standards.ieee.org/announcements/ 80211gapp2.html CA Collision Avoidance

41 WLAN Operator Cases

RIKU HONKANEN HELSINKI UNIVERSITY OF TECHNOLOGY [email protected]

ABSTRACT 1.2. Objective and Methods Many operators have been offering Wireless Local Area The objective of this report is to give an outlook to the Network (WLAN) services for few years now. There has WLAN operator business opportunities. We will start this been so far only little discussion about the success of these with an analysis of the environment and technology from operations and at the point of uncertainty about the third a business point of view. Later we will examine four dif- generation mobile phone techniques also WLAN has to ferent case studies: three Finnish WLAN operators and be thought as an option. This report gives an overview of for comparison one foreign operator. The scope of this the WLAN business environment and concentrates on the project is in some parts limited only to the Finnish market history and current situation of three Finnish operators situation to outline the topic from getting too large. The and one foreign operator in order to evaluate the success report will discuss the technical issues of WLAN tech- of their WLAN services. Two of these operators have nology only when it is important in order to understand failed in their WLAN operations in some way and the the business view. other two are making low returns. The operators cannot The research that I have done for this report is based on be encouraged to go to the WLAN service business published material like articles, press releases and opera- strongly because it is not currently responding to the cus- tors’ announcements. Because this field of study is evolv- tomer needs. ing rapidly at the moment, I found it more useful to base my sources on articles that can give more recent information than books. The usage of articles is also a good way 1 INTRODUCTION to get an overall picture of the situation because there are Some operators are currently offering Wireless Local Area so many different aspects from various writers. Network (WLAN) services. The media and operators are concentrating on the next generation mobile phone techniques but WLAN technique should be studied from the 2 EXTERNAL ENVIRONMENT ANALYSIS operator point of view to understand what the real poten- After the Internet and mobile communications revolutions tial of this technique is. This report was written for the there might now be a need for wireless data solutions in Research Seminar on Telecommunication Business II, the market. Different techniques have been introduced to which was held in the Helsinki University of Technology become a de facto standard but so far we are in the early at the department of Computer Science in the spring 2003. stage of wireless data communications and without a clear The theme of this year’s seminar is wireless technologies view of what technique is going to be the “killer app” the and this report concentrates on the business opportunities market is acting carefully. Analysis of the external envi- and case studies of Wireless Local Area Networks. ronment in Finland will be done in the following subsec- tions to get an overview of the basis for WLAN business. 1.1. Background Wireless Local Area Network is a technology for con- 2.1. General Environment necting devices through wireless connection to a Local Although the Finnish population is small in size and econo- Area Network. WLAN is also known by the name of stan- mies of scale cannot take off in Finland, the Finns have dard IEEE 802.11b and sometimes called as Wi-Fi (short proved to be early adaptors for modern digital devices for Wireless Fidelity). To be more specific WLAN is a like mobile phones. In fact Finland and Sweden are al- broader concept and Wi-fi (same as 802.11b) is one spe- ready one of the fastest adopters of WLAN compared to cific standard implementation of WLAN. The IEEE the rest of the Europe. Besides from the few bigger cities, 802.11 standard was published in 1999 and it has been Finnish population is very scattered and the wireless com- gradually adopted in building modern networks since then. munications solutions have importance especially in the WLAN network can be build with a base station, which is outlying districts where distances are long and building possibly an access point to another network, and devices new cable lines for a small amount of households is ex- with WLAN interfaces that are connected to the base sta- pensive. tion with radio signals.

42 The Finns have good purchasing power on average and 3 TECHNOLOGY ANALYSIS they are probably ready to pay for a wireless data tech- Originally the WLAN technology was designed to create nique if they find the price low enough and the speed of wireless LANs and LANs interoperating with some wire- the connection is rather fast. At the moment the Finnish less devices. From the operator point of view it is seen as economy is not at strongest possible stage and many in- a way of offering wireless Internet connections or other dustries are prudent about the current investments. The wireless data services. Although the bit transfer ratios of general economical situation reflects on the minds of the Wi-Fi are fast one base station only covers at maximum people and they need really value creating services be- about the same area than the traditional local area net- fore they will adopt new technologies. The media have a works. Therefore the technology is situated somewhere great influence on attitudes and if some technique is seen between the UMTS and high speed wireless technologies as a flop in the media it will probably come true as the like HiperLAN in the bit rate/coverage comparison as can buyers are looking for products that can create networks be seen from figure 3-1 below. of many users.

2.2. Industry Environment The telecommunications operator industry is now in a point where mobile communications and Internet operations are growing quite steadily and are already quite sta- bilized. The industry is waiting for the next area for making revenues with new operations like the wireless data services. There are already quite many existing telecommunications operators in Finland which all might have resources and willingness to take action to create new wireless data services. Although some operators are already offering these services in some form, the industry is quite open for new entrants as has been seen by the example of new operator DNA Finland. None of the current operators has a large installed base of customers in WLAN solutions. Figure 3-1. Situating wireless technologies. (Siitonen If WLAN will become successful among the customers, 2001) there may be a sudden and intense rivalry between the operators. If it gets to this point, Sonera and Radiolinja will probably be strong in the market but there is also an 3.1. Costs opportunity for totally new entrants to base their opera- The current de facto WLAN works in the 2.4GHz band tions purely on WLAN services and build their whole and the next generation will be working in the 5GHz band. business concept around this to make life harder on the These frequency bands are license-exempt in most coun- bigger operators. This might lead into a situation where tries. In this sense the case of WLAN is very different new entrants are cutting the prices and especially Sonera from the Universal Mobile Telecommunications Service is trying to aim its products to its large installed base of (UMTS) licenses, which in many countries demanded the current other services. huge investments from the operators. European operators The WLAN business problems are not however a conse- invested 150 billion euros on the third generation mobile quence of possible rivalry between the operators, but it is network licenses in the year 2000 (Rainisto 2002). For the threat of substitute products. It is totally unpredict- example Sonera paid 3.5 billion Euros for the German able at the moment if the operators’ customers will ever licenses for the next 20 years, which concludes to 0.5 use the WLAN in a large scale. There are already com- million euros per day even without the loan interests peting techniques available that might become substitutes (Laitila 2001). From this perspective the WLAN is defi- for WLAN. Sonera and Telia, currently the biggest Finn- nitely much more cost efficient than 3G mobile technolo- ish WLAN operators, believe that this year will be a break- gies. through for WLAN as business individuals will buy more To set up a WLAN network the operator will need base and more WLAN interface cards and WLAN roaming will stations located in visible locations. One base station cur- increase (Muukkonen 2003). Sonera is however offering rently costs about 1000 euros, which can be considered also the substitute products and tries to be competitive in affordable because one correctly situated base station every area of mobile communications and not focusing should easily be able to return this initial investment, and totally on any specific new technologies yet. After the year this is only a one tenth of the cost of a basic mobile base 2003 we may be a little bit wiser about which technolo- station (Woolcock et al. 2001). Finnish energy company gies will become more of a flop and which will begin the Vantaan Energia has recently decided to start building exponential growth of usage. WLAN networks for home subscribers in the Vantaa area. They are planning to build 115 base stations on roof tops

43 during the year 2003 and they have estimated the total services do not offer much extra value because the same costs for building the network to be 1,5 million euros and functions can be done using other devices and there is no the network should cover 33 500 households (Haukkasalo real killer application that would take advantage of the 2003). WLAN can be cost efficient in the densely popu- new UMTS possibilities. The buyers are suspicious after lated locations but building a WLAN network covering the UMTS license stir in the media and the cost of the large areas easily becomes a very expensive task. devices. But nevertheless the future of WLAN is sharing many of these same problems in its own way. 3.2. Competing Techniques 3.2.2. Fixed Media Broadband Access Techniques From the competing techniques point of view WLAN can be seen on two different aspects. On the other hand it is There is an obvious demand for broadband Internet ac- competing with traditional fixed line broadband access cess for home users at an affordable price. Currently in networks and on the other it competes with mobile phone Finland the most widely used solutions are ADSL and wireless multimedia and data solutions. It can also be seen cable modem techniques. The subscription fees of both as a question between indoors and outdoors usage, be- techniques are currently around 50 euros per month for cause fixed lines are clearly for indoors use and third gen- the end users. The most important issues for the home eration mobile phones are competing more in the outdoors customers are the ratio between price and speed of the environment. connection and also the reliability of the Internet connection. At the end of the year 2002 there were about 170 000 broadband Internet access subscriptions in Finland 3.2.1. Third Generation Mobile Networks (Sjöström 2002) and the figures seem to be rising still Like the second-generation mobile networks, the third quite rapidly. generation (3G) is basically a mobile phone network, with WLAN can be considered as a respective competing tech- added intelligence and higher data speeds. The most prom- nology for the current broadband access techniques be- ising standard seems to be the Universal Mobile Commu- cause of its speed and price. The Finnish energy com- nications Service (UMTS), which is already in small use pany Vantaan Energia has considered offering power line in some cities in Finland. So far one of the biggest prob- communications, but recently they announced that instead lems with UMTS has been that many operators had to they will build a WLAN broadband access network for pay huge amounts of money to get the licenses for differ- households. They are going to charge less than 50 euros ent countries and after the investments in licenses the per month for the subscription and they believe that they operators could not afford to build the actual network yet. will generate 2.5 million annual revenues after few years In Finland the situation is a little bit different because the (Haukkasalo 2003). The question is: will they succeed in Finnish government gave the licenses for free to advance what others have already failed? We will discuss this is- the building of 3G networks. Still, it is at the moment sue more in our case studies. WLAN can be offered at a very uncertain in Finland if the 3G networks will ever lower price than for example ADSL in the urban areas but become notable. the reliability of the access can be questionable because Despite the license costs and the currently unclear future, the quality of the access is affected by several influences UMTS has strong features that can be said for its advan- like the weather or interfering devices. tage. UMTS is a standard that has been endorsed by major standards bodies and manufacturers of the mobile phone markets. It is based on the already successful Glo- 4 CASE STUDIES bal System for Mobile Communication (GSM) standard Now we will go through four different case studies: Jippii and therefore supports for example roaming. It is capable Group Oyj, Sonera Oyj, Telia and MobileStar Inc. In our of 2 Mbps data transfer rate in theory and has coverage first case study we will look into the operations of Finn- area of over 400 meters from the base station, which ish WLAN pioneer which eventually failed in the WLAN should work more reliably over obstacles than WLAN. sector. Sonera and Telia are now merged but they are at Very important issue to note is also that the licensing of the moment both providing separate WLAN services in the 3G frequencies prevents other devices to use the same Finland and the final case of MobileStar is an interesting frequency band. With WLAN it is possible that micro- example of a foreign operator with WLAN services as wave ovens or any other devices may be using the same their core competence area. frequency than one’s network connection and interferes with it. In the case of Finland where the licenses were there for the taking, this reduces the technical concerns of 4.1. Jippii Group Oyj (formerly and currently UMTS compared to WLAN technique. Saunalahti) 3G networks do compete with WLAN in the markets of mobile users like businessmen with laptops, mobile phones 4.1.1. Beginning of Jippii Freedom Service and handheld devices. WLAN networks are much more Jippii Group, formerly concentrated in being an Internet cost efficient to build but 3G networks seem to be more service provider, went through big changes during the year reliable and have better coverage areas. Current UMTS 2000. First of all it got listed on the Helsinki Stock Ex-

44 change (HEX) and the company redefined its business to not believe that they should invest in service, which had be no longer only an ISP but a multi service provider. only about 700 home customers. 700 customers would Based on this decision Jippii concentrated more on mo- deliver only a maximum of 400 000 euros annual income bile entertainment, GSM and WLAN networks. Jippii which could just barely cover the salaries of the employ- Group launched Jippii Freedom, a broadband WLAN ees of WNS. Internet access service, during the year 2000 in few cities The base stations that WNS was using were not designed in Finland. (Jippii Group 2000) for outdoor use at all but for indoor office usage (Ahokas Jippii had the first mover advantage in this because they 2002). Therefore the connections between the base sta- were offering WLAN based Internet connection for home tions and the customers had high error rates. Buildings users as an alternative to different DSL and cable modem and other obstacles between the base station and custom- techniques. At that time for example ADSL connections ers made usage of the service impossible also somewhere were costing more than the Jippii Freedom service. Ev- in the promised coverage areas. WNS employees could erything looked promising as they started to do the pio- not keep up with the needed maintenance of the base sta- neering work for large WLAN Internet access networks tions and they were fighting a losing battle. in Finland. The company saw the WLAN technique as an opportunity for all the 2G mobile operators, who had not bought the expensive UMTS licenses, to provide wire- 4.1.3. Current Situation less data networks. WNS Oy still exists as a wholly owned subsidiary of Jippii In February 2001 WLAN services were spun off from the Group but it has no employees and it no more sells Jippii Jippii Group into a new subsidiary company named WNS Freedom service. At the end of the year 2002 Jippii Group Oy (Wireless Network Services), which immediately announced that they will no longer charge anything for started to raise funds aiming at 17 million euros. WNS the existing Jippii Freedom customers because they can- intended to use these funds for building WLAN networks not maintain the base stations anymore. Currently there in urban areas of Finland, Germany and England. This are about 400 signed customers still using the service. money was to be invested in thousands of new base sta- There are no more foreign WLAN operations by Jippii tions (Jippii Group 2001). WNS was supposed to make Group. Jippii will no longer offer any WLAN services deals with several foreign operators for delivering their and instead concentrates on fixed line Internet service WLAN operations and they estimated to break even with providing (Ahokas 2002). the costs already during 2002 and the revenues of WNS The approach that Jippii took towards WLAN was that was supposed to be 80 million euros already in 2003 WLAN was just a new technology to use for their tradi- (Holtari 2001). tional ISP business. They wanted to build large networks By the end of June 2001 Jippii Freedom had 700 signed covering the biggest cities to provide home users a fast customers in Finland and at that time Jippii Group an- lane to the Internet. It was seen as a cheap technology to nounced that it would continue providing WLAN con- provide inexpensive Internet connections but their own nections in Europe starting from Switzerland during the equipment was inadequate for outdoors use and they never year 2001 (Jippii Group 2001). The amount of customers realized that the initial costs of the customers caused high was actually extremely low for already expanding the switching costs because customers had to invest at least business to foreign countries before really seeing how it 500 euros for their WLAN equipment to try the Jippii will take off, but people at WNS Oy thought that they had Freedom. to be fast and offer the same concept in other European countries immediately. 4.2. Sonera Oyj

4.1.2. What Happened 4.2.1. Beginning of wGate Service There are probably many reasons for what eventually did Sonera launched the service called Sonera wGate (Wire- happen. Some of those reasons are that WNS never got less Gate) in June 2000. Sonera’s strategy was different their 17 million euros, they had many problems with the from Jippii Group’s approach right from the beginning. technology they were using and sales figures of the Jippii Instead of trying to market wGate as a wireless broad- Freedom service were really low. With all these problems band access for home users, they targeted the wGate to the Jippii Freedom was doomed to vanish from the mar- be a complementary product for business users (Salz- ket. The best achievement of Jippii Freedom was to be Trautman 2000). They already had an installed base of only 0.04% of the whole annual revenues of Jippii Group mobile phone business customers that were using some Oyj (Ahokas 2002). mobile data services. By people at Sonera it was seen that The investors lost their faith in Jippii Group as for the WLAN has its demand in the public areas like hotels, air- total multi operation view because the company could not ports, congress areas and wherever the customers have deliver the needed profits from the new operations and at the actual need for the Internet but their regular office the same time the whole IT business world collapsed as network is unreachable. Sonera wanted to add value for the great bubble of big promises burst. Investors could their existing wireless operations and already in the year 2000 they thought wGate would be usable together with

45 the upcoming 3G networks. Sonera took Nokia Networks 4.3.2. Current situation as a partner in developing their WLAN products. As Nokia Telia has been able to carry out the plans that they made offered them the hardware solutions for the operator side few years ago. Although nothing has happened at a very of broadband WLAN access they could concentrate in fast pace, it has been a steady growth with gradually get- building the actual hotspots. In June 2001, with both ting contracts with hotel chains and as well getting part- Sonera and Nokia being competent in the GSM networks, ner operators for roaming. HomeRun can at the moment they based their WLAN solution on SIM card based au- be used via Telia’s own hotspots in Sweden, Finland, thentication to guarantee security for the otherwise quite Denmark and Norway and by roaming in France, Italy, insecure WLAN standard (Salz- Trautman 2000). Around Belgium, Thailand and the U.S. In Finland they are natu- the same time Sonera made a contract with the Finnish rally using roaming with all Sonera’s wGate hotspots and Civil Aviation Administration about equipping all the 25 in addition they have about 40 hotspots of their own. In Finnish airports with WLAN hotspots. Sweden they have about 400 sites in many different cities. They have introduced new hotspots steadily varying 4.2.2. Current Situation from 1 to 10 sites per month. Like Sonera, Telia has not been pushing the WLAN technology aggressively, but is Sonera wGate has not grown into a large number of gathering information on delivering broadband wireless hotspots. In 2002 Sonera made a contract with Sokos data services for the business customers. Hotels about equipping 17 hotels with wGate hotspots. Sonera currently has hotspots in Finnish airports, hotels, congress areas and company premises with altogether 4.4. MobileStar (currently T-Mobile Hotspot) about 160 hotspots. All of these hotspots are located in Finland and there is no evidence that this would change. The merger with the Swedish operator Telia has changed 4.4.1. Rise of MobileStar the situation a bit so that Sonera wGate now has a roam- MobileStar was an American, Texas-based, company ing agreement with Telia’s HomeRun. founded in 1996. It was the world’s first company to start The core competence of Sonera is being a mobile opera- building a larger network of broadband wireless access tor. The way they conduct their WLAN business is based with the hotspot approach. This actually was their busi- on this same point of view and it is an additive service for ness idea and core competence, so in this way they differ their other major mobile services. WLAN is currently very from our Finnish case companies. The company received small in their overall portfolio and they keep it very steady its funding from investors such as Mayfield Fund, Blue- and wait and see what is going to be the future of WLAN print Ventures, Norwest Venture Partners, Sienna Hold- and other emerging new technologies. They are keeping ings and Texas Pacific Group (IBM 2001). At first they this service alive with a low profile but probably want to concentrated on hotels and in the year 2000, 100 from keep evolving it slowly to be able to scale it up if the total of 130 sites were hotels in major cities of US. In WLAN market is going to grow. their marketing MobileStar emphasized the ability to use virtual private network to connect to the customer’s corporate intranet and made clear that MobileStar had the 4.3. Telia biggest wireless network for their customers with secure technology. 4.3.1 Beginning of Telia HomeRun Service In the year 2001 they made IBM their partner. IBM’s role was to provide site surveys, equipment build and deploy- Telia HomeRun Service was launched already in Septem- ment, and cabling and wiring. At this point MobileStar ber 1999. It was available through hotspots in Stockholm’s had about 170 hotspots deployed around the United States main airport and larger railway stations (Computerwire but planned to have 2000 of them by the end of the year Inc. 1999). Telia made a contract with the Scandinavian (IBM 2001). MobileStar also got its biggest single cus- airline company SAS for deploying their airport lounges tomer, Starbucks, in January 2001 by signing a contract with HomeRun hotspots and SAS was already talking about equipping more than 4000 Starbucks coffee shops about in-flight WLAN solutions at that point. with wireless network within two years (Arensman 2001). In April 2000 Telia had about 50 hotspots in Sweden and MobileStar was growing really fast (about 250% annual they were planning to go international first with HomeRun growth) and it looked like the wireless hotspots had fi- starting from neighboring countries Denmark, Finland and nally taken off after five years of their efforts. Norway. Secondly Telia was looking for partners in other European countries in order to get roaming for their ser- 4.4.2. Fall of MobileStar vice in the regions where they did not operate themselves (Uimonen 2000). Telia was aiming at traveling business All in the sudden, in October 2001, MobileStar announced customers in Sweden but also the ones who traveled within that it had lain off nearly its entire staff. It also hired the Europe. They realized very early that roaming was a big Diablo Management Group to oversee an orderly sale of obstacle for the growth of the WLAN operator services. the assets. The reason for this was that MobileStar had tried to raise funds for at least six months before this and

46 still few weeks earlier they had a promising deal coming up but at the last minute they failed to sign that deal. This was a serious blow to the whole nascent WLAN community because MobileStar was definitely one of the highest-profile companies in this market (Charny 2001).

4.4.3. Current situation Few weeks after the sudden demise of MobileStar, the CEO Robert Kaiser gave an interview saying that they have no plans to shut down their existing network and this has proven to be true because at the moment the network is running. At the end of October MobileStar was applied to be acquired by the US mobile operator VoiceStream that at that time was already a part of the German T-Mobile operator. VoiceStream still had faith in the WLAN business and they got the existing MobileStar Figure 4-1, Operator strategies. Failed strategies services up and running again. Currently marked with red colour www.mobilestar.com is redirected to www.t-mobile.com/ hotspot, and the new name of the service is T-Mobile Common feature for both failure cases is the driving of Hotspot. They now have almost 2300 hotspots around the fast expansion. This approach demands a lot of funding US and they recently signed a deal for equipping 400 and is based on the view that large network will attract Borders Books and Music sites (T-Mobile Int. 2002). the customers eventually and the company can slow down the expansion gradually. Both Jippii and MobileStar failed Having a big operator like T-Mobile running the WLAN to get the support of investors and all the cash was flow- business it seems like a lot more promising service than ing out of the companies. Yet another mistake for Jippii MobileStar never really was on its own. T-Mobile gives was to target their product for home users who were en- much reliability for the Hotspot operations even though countering big initial costs of the equipment and still the they are getting their profits from their new GSM net- service was unreliable even with outdoors antennas. work customers. In any case they have the biggest single operator Wi-Fi network in the world. MobileStar may have The more successful cases Sonera and Telia can by no gone bankrupt because it was expanding too fast with low means be considered as successful. Their WLAN opera- profits and counting on investors and on the other hand tions have been and are more like test laboratories for the they never got enough customers with existing hotspots companies, which are getting all the profits from other to keep the return on investment high. services like the GSM. Their effort seems to be so slow that at the moment it is going nowhere. Still, they might be on a striking distance if Wi-Fi is getting more popular. 4.5. Case Studies Summary In my opinion the most prominent case is the T-Mobile Hotspot, which has risen from the fallout of MobileStar’s Each case shows that WLAN market has not yet been a bankruptcy. They have a strong company in other areas place to make good profits. The only company that tried and they seem to be willing to expand their service. They to do WLAN its core business went bankrupt and was are pretty much dominating the Wi-Fi hotspot markets in acquired by a big operator. Jippii’s WLAN operations the US. If there will be a strong competitor for them in became a total failure that has now been buried almost those markets, it would be an excellent opportunity for totally. The figure 4-1 shows the different approaches the both companies to sign a roaming contract and truly taken by all the case companies. It is a rough division of get a good coverage for both companies’ services. This strategies in three central strategic areas: the area of op- might be a breaking point for getting the customer flows eration, how fast the company is expanding the service to an exponential growth. and what is the objective of network’s coverage.

5. Problems As can be concluded from the case studies already, WLAN market area is not easy. There are several problems to be solved in order to make this a lucrative industry. Some of the biggest problems are discussed in the following sub- sections.

5.1. Lack of Roaming 200, 400 or 2000 hotspots may first sound like a good amount of sites, but the fact is that if we scatter T-Mobile

47 Hotspot’s over 2000 hotspots around the US ground, it is We must look at the indicators given by similar services: the same as having no coverage at all. Although the sites people are willing to pay for a broadband access to the have been picked out so that they could best reach the Internet at home but the adaptation of mobile data ser- target customers, it is impossible to give good service for vices has been slow in every sector. This implicates that these customers with only your own hotspots. If a busi- there is a bigger demand for high-speed data than wire- nessman is using Sonera wGate at the Helsinki airport less data. People are currently using actively the fixed and flies to London, he cannot use the wGate from the line Internet access at home and at work. The reasons that moment he leaves the airport lounge in Helsinki. they would want to take WLAN service instead of traditional xDSL service are based on the cost, speed and reli- Roaming is definitely needed for public commercial ability of the service but there is hardly any need to have WLAN services to ever become widely used and this is a wireless connection at home or at work because the fixed also realized by many of the operators. So far there are so lines work quite well for even laptop computers. few roaming contracts between the operators that they do not add enough value from the customer point of view. The hotspot approach instead of wide coverage is a good Roaming should be as global as it is with GSM to attract choice as long as we are thinking about laptops instead of the customers and making the services truly global. Be- handheld devices. Laptop’s usage area is much more re- cause many of the bigger WLAN operators are also GSM stricted because it requires more physical space and it is operators it should be an advantage for them to expand quite impossible to use it while walking for example. Quite the existing GSM roaming deals to include Wi-Fi roam- many people use their laptops at home more than on the ing as well. road, so then you would need a different kind of connection because hotspots are not placed usually on residen- Roaming is needed to create value for the WLAN net- tial areas. Many people would even choose GSM based work. If one can use the same service everywhere in the modem with very low speed over the WLAN because world or only at the airports, there is a huge difference in GSM works almost everywhere and almost every time. the value of the service. A network truly covering the everyday life locations might be much more attractive for Customer needs are a very big issue in the future of wire- customers than the current situation. less data. If there truly is a need for this type of service then one of the mobile data techniques will probably be a killer app but which one depends once again on customer 5.2. Customer Needs needs. There may be different kinds of target customers Many people within the telecommunications industry see for each technology but probably some service will be mobile data services as a future trend as for example preferred over the others because there is a need for con- Nokia’s view in figure 5-1. At the moment operators are sistent technology. Without consistency every technique making money with the voice delivery markets but adop- will fail in some way because customers need the net- tion of the existing mobile data services has been slow. work effects that can only be reached by using the same Some people believe this is because the speed is too low technique and therefore creating the network of users for in the existing services and the reliability of the service is that specific technique. too low. There may be a seed of truth in these views but it can also been stated that services described in the case 5.3. Billing studies have neither taken off although the speed of the connection is already rather high. This may be a question Billing is a tricky subject for all mobile data services. It is of low availability of these services as was discussed in a different kind of service than mobile voice stream and the previous subsection. the operator should find the right way to price the usage either by using a flat monthly fee, base it on the sent data or base it on the time of the connection. The even more important issue is how to handle billing when roaming because different operators choose different pricing approaches. This issue can be solved by making agreements between the operators about how they will charge each other for the visiting customers.

5.4. Security Issues Security has been a big concern with WLAN right from the start. Like in every wireless device, eavesdropping is much easier than with cables and with the ranges of WLAN one might be able to do eavesdropping even from miles away. Because eavesdropping is easy, the use of WLAN requires encryption and authentication methods to be used Figure 5-1 (Nokia 2001) in commercial applications. This is well realized in the business and techniques like virtual private networks

48 (VPN) use tunneling and encryption to communicate with operations but one has to get many customers locked in private networks over the Internet. This problem can be to do it, and nobody seems to know how to get these cus- solved with technical solutions, but the WLAN technique tomers coming. has an association to insecurity in many people’s minds. Most promising approach for utilizing WLAN from the The IEEE 802.11 standard includes an encryption scheme operator perspective is to combine these services with called Wired Equivalent Privacy (WEP), but this encryp- other data services. If customers would be able to use for tion is too easy to break and therefore it is not advisable example UMTS and WLAN through a bundled contract to use WEP in commercial operator services. it would open whole new opportunities. Customers would get eventually a full coverage network working every- 5.5. Obstacles where and the WLAN hotspots would boost the connection whenever customer enters one of these areas. It would WLAN technologies have a need for a clear visibility be a WLAN service, which would use slower connec- between the base station and the terminal. For example in tions wherever the WLAN is not available and this would the Jippii’s case of providing Internet for home users, the guarantee the users that they could access the Internet or obstacles between the base stations and the terminals di- their corporate intranets anywhere. minished much of the coverage of the base stations and led to complaining customers. Visibility is an important issue in urban areas where many base stations are situ- 7 CONCLUSIONS ated because those areas are commonly full of obstacles like large buildings. It is a bit easier to use WLAN in- The companies that were examined in the case studies doors and in open spaces to get better connection. In out- have either failed in their WLAN operations or they are doors use the connection might fail even when the weather making low returns. WLAN looks promising technique is foggy or rainy. Like the case of security this is more a and will grow, but it is not coming to be a big hit for the technical problem than a business-wise problem. operators. It can be best used as a complementary product for the existing and upcoming mobile data technologies like perhaps UMTS. The technology is cheap to 6 FUTURE AND REVENUE implement and there are no license costs but there are OPPORTUNITIES neither enough customers. For some reason WLAN is not answering to the customer needs of the operators but con- The future of WLAN operator business is very much un- tinues to increase in private networks. certain. So far these services have created only very small revenues and some operators have failed totally in this The lack of roaming is the biggest problem for WLAN at market already at this point. There are a number of com- the moment and the future does not look promising. From peting technologies striving for the same market area and the operator point of view this technique remains on a the future might include only one of these or a combina- good to know basis so far, because the markets for public tion of many techniques. access networks are small and the initial costs are too high for the home users to use WLAN for their home Internet The size of the customer base is clearly dictated by the access. The current services do not meet the customer number of business people carrying wirelessly enabled needs at the needed level. mobile PC devices at the moment. Intel has recently in- formed that it will include a WLAN chip to every mobile processor in near future. This means that there will be REFERENCES three out of four laptops having WLAN capabilities within few years. It is very probable that the WLAN will be- Ahokas, K., 2002. Jippii luopuu langattomasta. Tietoviikko. come a popular technology for laptop users and also for August 8. some desktop computers for creating networks but is very Arensman, R., 2001. Cutting the Cord. Electronic Business. uncertain if there is going to be revenues for operators June. building public access networks because the companies Charny, B., 2001. MobileStar shutting down network. CNET can build small networks for themselves, connect these to News.Com. October 11. Found at: http://news.com.com/2100- the existing wired network and connect to the Internet via 1033-274288.html?legacy=cnet [referenced March 14, 2003] wired network if needed. Computerwire Inc. 1999. Telia Unveils Wireless Internet The current operator services are priced at about 30 to 50 Access Service. Computergram International. October 4. euros a month with the flat rates. The base station prices Haukkasalo, A., 2003. Datasähkötekniikka kypsyy käyttöön. are around 1000 euros, so we could estimate that about Tekniikka & Talous. January 30. less than 10 customers could return the needed money to build one hotspot in one year. Because of the low price of Holtari, S., 2001. Järki valtaa teleyhtiöt. Talouselämä. May 4. the base stations this is an attractive notion, because the IBM Corp. 2001. MobileStar Selects IBM to Grow Broadband Wireless Network. Press release. April 11. Found at: http://www- return on investment looks quite high. The only problem 916.ibm.com/press/prnews.nsf/jan/C625220EFA4 is that by now the services have not attracted the custom- 0E77385256A2B004C547F [referenced March 14, 2003] ers. There is a good chance of making profits with these Jippii Group, 2000. Financial statement notice 1.1.2000-

49 31.12.2000 Siitonen, A., 2001. WLAN. Found at: http://www.tml.hut.fi/ Opinnot/T- 110.300/2002/Luennot/l1120a.pdf [referenced on Jippii Group, 2001. Interim report January-June February 23, 2003] Laitila, M., 2001. Jippii haastaa Elisan. Talouselämä. May 4. Sjöström, M., 2002. Yksityiskäyttäjän kaista levenee hitaasti. Muukkonen, H., 2003. Wlan kypsyy käyttöön. Tietoviikko. MikroPC. August 29. T-Mobile International, 2002. January 31. T-Mobile International Reports Third Quarter 2002 Results Of Nokia, 2001. Nokia in Messaging. White paper. Found at: http:/ U.S. Operations. Financial release. November 14. Found at: /www.nokia.com/downloads/aboutnokia/press/pdf/NIM.pdf http://www.t-mobile.com/company/investors/financial_releases/ [referenced March 19, 2003] 2002_1114.asp [referenced March 14, 2003] Rainisto, S., 2002. Kallis pakkomielle. Talouselämä. Uimonen, T., 2000. Symbol, Telia make a deal. InfoWorld. September 27. April 3. Salz-Trautman, P., 2000. Operators hit the WLAN hotspots. Woolcock, K., Murphy, S., Wyatt, P., Tyndall, M., 2001. The Communications Week International. August 14. Barbarians at the Gate – Wireless LAN Storms 3G Citadel. Nomura. March 15.

50 Ad Hoc Networking as an Internet Access Technology

HENRIK PETANDER OLLI SAVOLAINEN HELSINKI UNIVERSITY OF TECHNOLOGY HELSINKI UNIVERSITY OF TECHNOLOGY [email protected] [email protected]

ABSTRACT tives and possible models of arranging Internet connectivity. After the technological background is given, we Ad hoc networks are self-organizing wireless networks, give a general view on applications of ad hoc networks. in which also end nodes act as routers. Ad hoc network- Finally, we analyze the timing and markets of using ad ing improves the efficiency and range of fixed and mo- hoc networks for connecting to Internet and conclude our bile Internet access and enables totally new applications study. such as sensor networks. Advances in radio technology and ad hoc routing protocols are needed for wide spread use of ad hoc networking. These technologies also need to be standardized to offer a sustainable basis for 2 TECHNOLOGY BACKGROUND interoperating products and a wider market. In this paper An ad-hoc wireless network is a multi-hop network with- we aim to give an overview on the technology of ad hoc out any base stations, or infrastructure. Ad hoc networks networking and its application to Internet access. support anytime and anywhere computing, allowing the spontaneous formation and deformation of mobile networks. A Mobile Ad hoc Network (MANET) consists of 1 INTRODUCTION a set of mobile hosts within communication range. These hosts are called nodes in MANET terminology. Wireless networking has several differences compared to wire-based networking. Being wireless means that there MANETs have several characteristics that have to be iden- is no need for a physical connection to the socket in the tified when discussing the underlying technology of ad wall, because data is transferred through air in form of hoc networks. (Corson, Macker, 1999) radio waves. Wireless networking also enables user mo- • Network topology is dynamic. Nodes are free to move bility. Basically, the chance to move freely anywhere, arbitrarily, which means that the topology of the network anytime, and still have a change to be connected to the may change randomly and rapidly. network, is a natural source of interest. The costs of building a wireless infrastructure are virtually nothing com- • Network is bandwidth-constrained and capacity of the pared to the costs of wire-based network installation. links may vary. All that is needed for wireless communication between • Operations are energy-constrained. Most nodes in a two computers is devices for transmitting and receiving MANET are usually running on batteries or on other ex- radio signals and the transmission medium for the sig- haustible means. nals, which in this case is air. There would not be a need • The physical security is limited. Ad hoc networks are for established network infrastructure, whatsoever. There- generally more prone to physical security threats than fore, a network could actually be formed anywhere and fixed-cable networks. anytime. Anyone could join the network while passing through, or otherwise coming to the radio transmission Some issues on the underlying technology of ad hoc net- range of the other computers. This kind of networking is works are considered as an introduction to the subject of called ad hoc networking. Ad hoc networks are self-orga- this paper. nizing wireless networks, in which also end nodes act as routers. Ad hoc networking allows nodes to communi- 2.1 Ad Hoc Routing cate over multiple wireless hops and form self-organizing networks. Each node in an ad hoc network participates in forming the network topology. As there are no dedicated routers, Ad hoc networking is an attractive concept and has vari- each node is for its own part responsible for routing pack- ous possibilities for different kinds of applications. In this ets to and from other nodes. Basically the routing infra- paper the applicability of ad hoc networks for Internet structure is yet similar to the one of Internet. There are access is discussed. We start this by introducing the un- many different routing protocols that provide informa- derlying technologies and their limitations as well as dis- tion to forward packets to the next hop. In an ad hoc net- cussing some important issues like participation incen- work it is necessary to manage topology changes, and all

51 the nodes are required to run routing protocols. The rout- Concretely, there are three major security issues related ing protocols used in Internet are typically not applicable to ad hoc networks: Wireless links are easy to eavesdrop; to ad hoc networks as such. Routing is vulnerable, because every node acts as a router and can potentially enter false routing information into In general, mobility, dynamic topologies, and the con- network; Cryptographic schemes are hard to implement, straints of power and bandwidth in ad hoc wireless net- because the lack of support of infrastructure makes cen- works have given the guidelines for routing protocol de- tral authorities needed for key management problematic. velopment. As nodes in a MANET usually have to deal (Zhou, Haas, 1999) with limited power resources, it is suitable to develop protocols which need a minimal amount of information ex- changes, thus minimizing radio communication and also 2.4 Billing and Incentives power consumption. The most important routing protocols for ad hoc networks are Ad Hoc On-Demand Dis- Because ad hoc networks do not have an established in- tance Vector (AODV), Dynamic Source Routing (DSR), frastructure, there is no entity to collect payment for in- Optimized Link State Routing (OLSR), and Topology frastructure services. However, as all the nodes in a Dissemination Based Reverse-path Forwarding (TBRPF) MANET work as routers, there may arise a question: why protocols. Toh (2002) gives a good overview of routing should the traffic between other nodes be forwarded? In protocols and more detailed specifications can be obtained fact, forwarding data between other nodes has two disin- from IETF MANET working group. centives for an autonomous node: energy expenditure and possible delays for its own data. Not forwarding the data would, in turn, result in impossibility of forming ad hoc 2.2 Radio networks. In a multi-hop packet radio network, such as an ad hoc Ad hoc networks can be formed for many different rea- network, each station participates cooperatively in for- sons, and there are very different views and needs for ac- warding traffic between other stations. With multiple short tual billing or incentive considerations. Incentives and hops link quality improves thus improving data through- billing in ad hoc networks, in general, is a rather difficult put and stations can use less power for transmission and subject, and it would perhaps be more fruitful to consider receiving. There are also several differences in the access possible money flows in case of each application of ad medium, as with radio signals propagation, interference, hoc networks separately. Of course, there is a need to con- frequency band choices, and such things have to be consider general schemes for incentives, like Ad hoc Partici- sidered more carefully (Shepard, 1995). Also real sup- pation Economy (APE) (Fratkin, Liu, Vijayaraghavan, port of mobility and power consumption are important 2002), to provide the ways of implementing these things issues, when considering radio transmission. within different applications. In theory multi hop wireless radio networks provide more capacity than traditional radio networks (Shepard, 1995). 2.5 Internet Connectivity This is because of the lower transmission power needed for the shorter links and hence more efficient reuse of If a node in ad hoc network has connectivity to the global frequency spectrum. In practice, the scalability of current Internet, it can also offer connectivity to the other nodes. radio technologies is far from the theoretical optimum There are several ways to organize this from the network (Gupta, Gray, Kumar, 2001). infrastructure point of view. An Internet gateway can be considered advertising itself as a default router. This will Common radio technologies used for Ad hoc networking work if the other nodes in the ad hoc network can adapt to are IEEE 802.11 Wireless LAN protocol family, the being connected to their default router by a multi-hop Bluetooth, HiperLAN/2 and possibly in the near future path through other nodes. This idea contrasts with the tra- also UWB. These differ for example in terms of frequency ditional model of a default router. (Perkins, 2001) spectrum used, data rates, transmission range, and modularization and medium access control methods. Another idea is to consider the entire ad hoc network as a single hop from the point of view of Internet routing. (Sun, Belding-Royer, Perkins, 2002) This view is analogous to 2.3 Security the way BGP (Border Gateway Protocol) characterizes an entire administrative system (AS) as a single hop in its Ad hoc wireless networks do not have any predefined in- route advertisements. frastructure and all network services are generated on the fly. It is obvious that the lack of support from any under- One solution to finding the gateway is based on IPv6 lying infrastructure and the possibility of attacks on wire- anycast. All gateways could have an assigned anycast for less links make security an important issue in ad hoc net- MANET gateway in their interface. Nodes could try to works. It is challenging to achieve security within ad hoc find a route to the anycast address with normal route dis- networks basically due to following reasons: dynamic to- covery. pologies and membership, vulnerable wireless links, and Another problem is related to addressing. In addition to roaming in dangerous environment. (Vinayakray-Jani, gateway, nodes need a topologically correct IP address. 2002) (Topologically correct from Internet infrastructures point

52 of view, of course.) The IP address should also be glo- tion required for sending data when compared to sending bally unique for connecting to Internet. One possible so- the data directly between communication end points, i.e., lution to this is Mobile IP (Perkins, 2002). So, after a a mobile terminal and a wireless access point. The use of gateway is found, one way or another, the next step would ad hoc protocols allows the networks to be self- be to make the default router a foreign agent for Mobile configurable, decreasing the amount of configuration IP. Then every node in the ad hoc network can appear to needed to set up a network. These theoretical characteris- be accessible as if it were still located on its home net- tics make ad hoc networking a disruptive technology. work. Finding foreign agents would be based on agent However, in practice these advantages cannot be fully advertisements and route requests to an assigned multicast exploited due to limitations in radio technologies and rout- address for “all mobility agents”. ing protocols, but offer still notable benefits in certain application areas. 2.6 Standardization and Patents In mobile networking the computers, ad hoc network nodes, have limited computational and storage capabili- Standardization in the ad hoc networking field is done ties, and battery life is also limited. Spectral efficiency simultaneously on separate fronts, because there are many and communications overhead should be minimized for different technical fields concerned with ad hoc network- scalable and efficient operation. With fixed access, bat- ing. Radio technologies and routing can be considered tery life is not an issue, but spectral efficiency and mini- the most important areas of standardization in ad hoc net- mal communications overhead are critical for large scale working. Different radio technologies have been standard- deployment. ized and are standardized all time. At the same time routing protocols specifically for MANETs are introduced and In this section we will investigate the use of ad hoc net- also standardized. Security and Quality of Service solu- working for Internet access. We divide the access into tions are still far from standardization. two sections: fixed and mobile access. Fixed broadband access technologies and mobile data services overlap to Generally all the radio technologies introduced earlier in some degree, especially for portable access to the Internet. this paper are standards as such. IEEE is the standardiza- Still the two applications have at least for now separate tion organization behind the 802.11 protocol family, markets, e.g., home and corporate Internet access for desk- whereas HiperLAN/2, for example, is standardized by top computers using fiber, DSL and cable modems and ETSI. The situation with routing protocols at the moment on the other hand GPRS and the emerging 3G WAN and is, that IETF MANET Working Group has most of the IEEE 802.11 LAN mobile data services. As speed for routing protocols as drafts and DSR, OLSR, and TBRPF mobile data access increases, these two markets may are going to be published as RFCs. AODV has recently merge to some degree at least for the corporate segment, been published as experimental RFC. in which the cost is not as big an issue, as in the residen- All in all, the amount of different standards for similar tial segment. tasks is large, and is likely to be large also with future Mobile data services can be seen to include also vehicu- development. This may lead to incompatibilities in the lar networks where computers in vehicles communicate future or at least be confusing people and companies with other vehicles and also with computers located in working in ad hoc networking field. the Internet. However, the vehicle to vehicle communication distinguishes this application from normal mobile Internet access. 3 AD HOC NETWORKING AS AN INTERNET ACCESS TECHNOLOGY

3.1 Applications for Ad Hoc Networking Ad hoc networking protocols allow building of self-con- Fixed broadband figuring multi hop wireless networks. The concept in it- Mobile data access, including self is generic and can be used in several application ar- vehicle networks eas. Ad hoc networking is a critical enabling technology to some of the applications, such as sensor networks, where Size Network Sensor networks as others, e.g., fixed wireless broadband access networks can operate more efficiently using ad hoc networking pro- Home and tocols. personal networks In general the use of ad hoc also known as mesh networking or multi hop wireless networking increases the spec- Node mobility tral efficiency of communications, thus increasing the communications capacity of the network and allowing Figure 1 Applications for ad hoc networking. higher speeds for an individual user. At the same time use of multiple wireless hops decreases the power consump- Figure 1 classifies potential applications of ad hoc net-

53 working based on the mobility of the nodes in the net- band wireless access could possibly also be employed for work and the size of the network. The technological chal- mobile or at least portable Internet access. However, lenges become more demanding as these parameters in- mobile use leads to a frequently changing network topol- crease. ogy. Changes in the network topology pose challenges to any routing protocol used in the network. The amount of routing protocol related signalling makes large-scale flat 3.2 Fixed Broadband Wireless Mobile ad hoc networks impractical. To overcome this The characteristics of ad hoc networking make it suitable limitation, a clustered or hierarchical approach is needed. to building fixed wireless broadband networks. A multi The fixed wireless network would then form the back- hop wireless network has shorter wireless hops than a point bone network to which mobile nodes could attach through to multi point network. According to Shorecliff Commu- a small number of hops via other mobile nodes as shown nications (Wery et. al, 2002) this leads to increased data in Figure 2 (Mesh Networks, 2001). rates and decreased power consumption while allowing According to Mesh Networks (Mesh Networks, 2001), use of radio technologies requiring line-of sight between operators could benefit from use of ad hoc networking as the transmitter and the receiver. a part of their mobile access solution as the use of multi In fixed wireless networks the routing protocols do not hop wireless networking allows higher data throughput need to adjust as quickly to changes as in MANETs, where rates. Customers would also benefit from longer battery the topology is constantly changing. The routing proto- life. Mesh networks solution solves the incentive prob- cols can be tuned to minimize signalling overhead, thus lem by forcing the customers to let their terminals act as allowing building of large scale networks. routers. This is achieved by a proprietary solution implemented mostly in firmware making the ad hoc operations However, the use of ad hoc networks for fixed broad- transparent to the user. band access presents some challenges: • Multiple wireless hops increase the end to end latency, which is a limiting factor in real-time interactive services, e.g., voice. The latency depends much on the specifics of the radio technology used, though. For non-interactive data services this is not an issue. • Integration with back-end systems for billing and authentication requires new protocols as customers are not connected directly to ISP equipment. Several companies provide products (Nokia, 2001), (Ra- diant) for building fixed wireless networks that use ad hoc (or multi hop wireless) networking as their basis. From an operators point of view, fixed wireless based on ad Figure 2. Mesh Enabled Architecture. Groups can form hoc technology allows building of high bandwidth access ad hoc peer-to-peer networks anytime, anywhere. No networks with relatively low initial capital costs. Use of network infrastructure is needed. MeshNetworks users wireless access allows operators not owning to compete can hop through each other to become part of the with incumbent operator. Wireless may be exploited also group. Peer-to-peer networking is supported for both in areas without existing cable or wired telephone infra- fixed and mobile subscribers. structure.

3.3 Extending Cellular Mobile Access Networks 4 TECHNOLOGY BARRIERS AND TIMING Mobile ad hoc networking can be used for extending the coverage of a cellular mobile network. This allows mobile users to access the network even when they are out- There exist a number of technological barriers to adop- side the range of any base station. The use of multiple tion of ad hoc networking for providing fixed or mobile hops between a mobile node and a base station improves Internet access. the signal quality, which either increases the data rate or Radio and link layer technologies are critical to ad hoc decreases the required transmission power. networking. Currently deployed radio and link technolo- The cellular network in question can be, e.g., a CDMA or gies, such as IEEE 802.11 WLAN can not be utilized ef- an IEEE 802.11 WLAN network. For example an user fectively for large scale ad hoc networking due to surfing the web in an internet cafe with his laptop com- scalability and performance issues (Gupta, Gray, Kumar, puter and a WLAN card could also provide access to other 2001). On the other hand there exist technologies, such users outside the range of the cafeteria's WLAN access as HiperLAN/2, which are well suited to use in ad hoc point. networking, but are, at least for now, too expensive for wide spread use. The products and protocols developed for fixed broad-

54 Scalability of current routing protocols, AODV and DSR working will allow users to enjoy faster connection speeds in flat, i.e., non-hierarchical ad hoc networks limits the at a lower cost. Operators will be able to provide larger size of the network approximately to around 100 nodes. coverage and faster data rates at lower prices by extend- The scalability can be improved with clustering and hier- ing their networks with user terminals acting as wireless archical routing. routers. The lack of established standards or commonly adopted Advances in radio technology and ad hoc routing proto- de facto standards for quality of service and security cols are needed for wide spread use of Ad hoc network- mechanisms hinders the deployment of ad hoc networking. New radio and link control technologies need to be ing by network operators. Security and quality-of service designed to meet the requirements of Ad hoc networking, in ad hoc networking are still not even considered for stan- which differ from those set by cellular radio technolo- dardization in IETF as they are still in research phase. gies. These technologies also need to be standardized to offer a sustainable basis for interoperating products and a Energy constraints limit routing and radio technologies. wider market. Routing protocols have to minimize the number of routing information updates, which consume energy in form The challenges are being met by several start ups trying of radio transmission overhead. Also radio technologies to provide a new wireless access architecture to the fixed have to have limited power consumption to be applicable and mobile Internet access market currently dominated in MANETs. The power issue has also an effect on par- by operators struggling under large amounts of debt and ticipation incentives. The threshold for forwarding data by entrenched equipment vendors, such as Nokia, Ericsson between other nodes gets higher, because radio traffic and Motorola. Only the near term future will show us, consumes own power “in vain”. This may be even bigger whether ad hoc networking is ready for more than con- issue in case of Internet connectivity. necting near by devices for ad hoc communications. Figure 1 illustrates network size and node mobility for different applications and it also helps in estimating the timing of the applications. Ad hoc networking applications in the lower left corner are in our opinion likely to REFERENCES realize first and the ones in the upper right corner last. This development is estimated in Figure 3. Corson, S. & Macker, J. (1999) Mobile Ad hoc Network- ing (MANET): Routing Protocol Performance Issues and Evaluation Considerations RFC 2501, January, 1999, http://www.ietf.org/rfc/rfc2501.txt. Now 2004 2006 2010 ? Fratkin, E. Liu, Y. & Vijayaraghavan, V. (2002) Partici- pation Incentives for Ad hoc Networks Project report. University of California. Gupta, P. Gray, R. & Kumar, P. R. (2001) An Experimen- tal Scaling Law for Ad Hoc Networks University of Illi- nois at Urbana-Champaign. Sensor networks, Home networks Mobile access, Gupta, P. & Kumar, P. R. (2001) Towards an information Personal comm., Vehicle comm. Fixed wireless access, Portable access theory of large networks: An achievable rate region, in Military & emergency IEEE Int. Symp. Info. Theory, Washington DC, June 2001. Mesh Networks (2001) MEA product sheet, 2001, Figure 3. Timing of ad hoc applications. http://www.meshnetworks.com. Nokia (2001) Nokia RoofTop Wireless Routing, white paper 2001. 5 CONCLUSIONS Ogier, R. et al. (2003) Topology Dissemination Based on Ad hoc networking allows nodes to communicate over Reverse-Path Forwarding (TBRPF). Internet Draft, multiple wireless hops and form self-organizing networks. March 3, 2003, http://www.ietf.org/internet-drafts/draft- These two features make it an important technology, which ietf-manet-tbrpf-07.txt. may have a large effect on wireless communications in the near future. Ad hoc networking improves the efficiency Perkins, C. E. (2001) Ad Hoc Networking Boston, US. and range of fixed and mobile Internet access and en- Addison-Wesley. ables totally new applications such as sensor networks. Perkins, C.E. (2002) IP Mobility Support for IPv4, RFC Use of ad hoc networking will also make wireless net- 3344, August 2002, working more flexible, as users do not need to be within http://www.ietf.org/rfc/rfc3344.txt?number=3344 the direct reach of wireless access points to access the Internet. Pister, K. (2001) On the Limits and Applications of MEMS Sensor Networks, Defense Science Study Group report, More efficient use of radio spectrum through ad hoc net- Institute for Defense Analysis, Alexandria, VA, 2001,

55 http://www-bsac.eecs.berkeley.edu/~tparsons/ Toh, C.-K. (2002) Ad Hoc Mobile Wireless Networks: PisterPublications/2001/DSSG_Pister.pdf Protocols and Systems, pp. 27-77. New Jersey, US. Prentice Hall. Radiant Networks Meshworks product sheet, http://www.radiantnetworks.com Wery, R. & Kar, S. & Woodrum, J. (2002) When WiFi meets mesh, Shorecliff communiucations, 2002 Shepard, T. J. (1995) Decentralized Channel Manage- ment in Scalable Multihop Spread-Spectrum Packet Ra- Vinayakray-Jani, P. (2002) Security Within Ad hoc Net- dio Networks. Ph.D. thesis, Massachusetts Institute of works PAMPAS’02 Workshop on Requirements for Mo- Technology. bile Privacy & Security. Sept. 16-17 2002. University of London. Sun, Y. Belding-Royer, E. M. & Perkins, Charles E. (2002) Internet Connectivity for Ad Hoc Mobile Networks. In- Zhou, L. & Haas, Z. J. Securing Ad Hoc Networks, ternational Journal of Wireless Information Networks, published in IEEE Network journal, volume 13, pp. 24- Special Issue on Mobile Ad Hoc Networks. April 2002. 30, 1999. http://citeseer.nj.nec.com/zhou99securing.html

56 Market Strategy Evaluation of Ultra Wide Band Technology

ARI HAAPANIEMI OLLI-PEKKA ISOLA HELSINKI UNIVERSITY OF TECHNOLOGY HELSINKI UNIVERSITY OF TECHNOLOGY [email protected] [email protected]

ABSTRACT usage, but still leaving some regulatory issues open. The Ultra Wideband (UWB) is a promising new technol- What makes the UWB different from traditional radio fre- ogy for wireless communications. It is also known with quency carrier transmission is that instead of using con- various other names such as impulse radio, carrier free stant carrier wave, UWB typically uses extremely short radio and base-band radio. pulses in time domain for sending information over the air. Short pulses – from a few tens of picoseconds to a Due to short pulses used, UWB is practically capable to few nanoseconds long time periods – cause UWB trans- achieve data rates of several tens of megabits per second. mission to occupy several GHz in frequency domain, usu- For fixed point-to-point applications data rates of hun- ally more than 1.5 GHZ. Clearly, bandwidth is much dreds of megabits are achievable. Due to a simple nature greater than the bandwidths of traditional narrowband of UWB transmission, chipsets including UWB receiver technologies. Very short pulses sent with minimal delays and –transmitter are easy to implement making UWB a enable UWB to achieve data rates up to 1Gbps, in theory. very cheap technology. These features have placed UWB In practice, though, data rates of tens or hundreds of mega- in the front of the competition for new wireless technolo- bits may be considered more realistic. gies when consumer electronics markets are considered. In addition to high data rates, UWB devices are capable Despite of its clear advantages UWB has not gained popu- to operate on very low power consumption levels, which larity, yet. This is partly due to a fact that, up until now, can be regarded as a considerable advantage when regulatory bodies has restricted the use of the UWB to handheld devices are considered. Also, multipath propa- limited number of applications, which have been mainly gation, where delayed, reflected signals cause interfer- for non-public use only. In United States, Federal Com- ence at the receiver end, is not a problem with UWB, due munications Commission (FCC) has been the main regu- to a short pulse width of the UWB signal. latory body, which has been leading the standardization of the UWB into everyday transmission technology. Eu- Features mentioned above, especially immunity for ropean regulators, like European Telecommunications multipath interference and possibility to achieve high data Standards Institute (ETSI), are expected to monitor and rates, make UWB ideal for wireless local area network- follow the rules issued by the FCC. By the end of 2003, ing (WLAN). first consumer products, which conform the rules issued At the same time than UWB develops, AdHoc network- by the FCC and utilize the UWB, are expected to appear ing and software radio are growing trends. None of these on the market. themselves are services that consumers may utilize, but combination of them could provide extensive set of new services extending concept of mobility to totally new ar- INTRODUCTION eas. AdHoc networking concept allows devices to estab- Until now, UWB has been mainly used in devices intended lish communication links between compliant devices on- for authorities and military use only. Devices such as demand basis. Software radio on the other hand enables ground- and wall-penetrating radars have been used for easy and cheap way to configure radio interface to com- rescue- and military purposes. Excellent signal penetra- ply with sensed radio environment. The UWB could pro- tion properties make UWB well suited for that kind of vide fast and secure transmission capability for both of modern radar applications. Low power spectral density these technologies. makes signal very hard to detect providing good founda- For more information consult seminar proceedings tion for secure transmission. In the early stage, these UWB AdHoc, Technology Strategy Analysis (Petander, characteristics lead to military to regard UWB as very Savolainen) and Software Defined Radio, Technology attractive technology. As a result, UWB has been used by Strategy Analysis (Kuikkaniemi). the US military since 1960’s. The UWB technology does not come totally without prob- In 1998, the FCC recognized the importance of the UWB lems, though. The UWB’s extensive utilisation of frequen- technology and started to process regulatory review. In cies has raised some concerns whether the UWB will cause February of 2002, the FCC adopted formal rules for UWB interference on existing solutions operating on licensed

57 frequency bands. There have been fears that Bluetooth less or mobile technology. Which ones are those that and Global Positioning System (GPS) are in threat. As a complement it, then? result, emission power used by the UWB applications has GSM, GPRS, EDGE and UTMS have far greater cover- been restricted to few mW by the FCC, allowing UWB to age and they offer roaming that can still be considered achieve efficient indoor operating range of 10-20 meters. quite unique feature. They are targeted to true converging Given operating ranges make UWB suitable for indoor applications while UWB seems to have pure data back- use, but for outdoor use, emission powers may be inad- ground. At the same time they do not offer as high through- equate. put than UWB. Nevertheless, one of the major design goals of the UWB As we have learned, mobile or wireless systems are actu- has been to make emitted power amount per frequency ally a part of wired infrastructure and replace cables only band small enough that UWB can co-exist with other wire- in the last hop. Wired solutions are always required as less transmission systems in the same frequency. Low backbones except in the last hop. transmission power would enable UWB systems to co- exist with other systems, like cellular phones, GPS equip- Our opinion is that competitors will be Bluetooth, today’s ment and wireless network systems. Co-existence would Wireless LANs and HiperLAN2. Actually, HiperLAN2 further allow UWB to utilize possible unused frequency may not have any chances to challenge today’s Wireless channels existing between channels used by the present, LANs and UWB. IEEE 802.11 based wireless LANs have widely adopted standards. Low power consumption could extremely strong market position and they offer transi- also translate into a significant battery life extension mak- tion path similar to Ethernet. ing UWB a very tempting solution to be used in handheld Three possible scenarios can be outlined. We study them devices. For example, UWB transmitter uses only frac- via two significant segments, namely WPAN (Wireless tion of the power used by the Bluetooth transmitter: In Personal Area Network) and WLAN (Wireless Local Area order to deliver 1 Mbps with efficient operating range of Network). WPAN can be thought as personal appliance 10 meters Bluetooth transmitter requires average of 1 mW network that surrounds single human. WLAN usually of radio frequency (RF) power, whereas UWB transmit- consist of multiple computers and may have connection ter could provide same throughput and range characteris- to the Internet. tics using only 10 uW RF power (J. Foerster, E. Green, S. Somayazulu, D. Leeper, Ultra-Wideband Technology for UWB remains as marginal technology in our first and most Short- or Medium-Range Wireless Communications, pp modest scenario. Most market research companies are 6). betting superior future for UWB compared to this one. We are quite sure that today’s WLANs have so strong position that it is extremely difficult to replace them with WIRELESS ALTERNATIVES other technology. Why to implement yet another technology if existing one offers the most wanted features? The Existing IEEE 802.11 standards, like 802.11b and series of IEEE 802.11 specifications allow users to up- 802.11g, have dominating position on WLAN market. In grade and expand networks when needed. Bluetooth domi- order UWB to challenge their dominance, it has to pro- nates WPAN segment. UWB gains popularity in applica- vide some considerable additional advantages when com- tion areas where it's features offer unique advantage. These pared to 802.11 family of the standards. Security features are automobile collision detection systems and wall- could be such. Because UWB is inherently more secure through radars used by rescue teams. than any of the 802.11 technologies, it could offer solution for security problems, caused by the “openness” of The emerging market of HDTV replacements in US plays the wireless networks. Higher data rates provided by UWB an important role in the second scenario. We estimate that cannot be underestimated as competitive advantage in this one is most probable scenario. UWB performs well WLAN market. as cable replacement technology (replacing Firewire-connections between computers and video recorders and com- UWB may easier win competition within Wireless Pri- puters and USB devices). High Definition TV replace- vate Area Networks (WPANs). Bluetooth has been the ments are big opportunity to the manufacturers and Cable dominating standard within this networking area, provid- TV providers to bundle wireless circuits to the TV sets. ing efficient data rates of approximately 1 Mbps. The This may not be the development that consumers initially UWB achieves easily far greater data rates with less power demand but they may find out that setting up full home consumption. Especially smaller power consumption can theatre system is a whole lot easier when there is no ca- be regarded as major advantages when handheld WPAN bling work to do. UWB may also replace Bluetooth in devices are considered. WPAN segment but we think that IEEE 802.11 series of specifications has so strong market establishment that USB and IEEE 802.11 based systems will co-exist and may SCENARIOS blend with each other instead of direct replacement. As a background for following scenarios we have to dis- The last scenario is really bold. We have put a condition cuss about complementing and competing technologies. here. In order to penetrate the whole short-range data com- UWB technology will not compete with every other wire- munications segment, the seamless integration with up-

58 per layer protocols is needed including successes in ear- allow room for more technologies in the long run. lier scenarios. Actually, this scenario means that from user Much of the success of UWB depends on how transmis- point of view, growth path of existing WLANs continues sion capabilities are used in upper protocol layers. There but the underlying transmission system is replaced with is a strong tendency to use IEEE 802.3 framing and IP UWB. everywhere. If IEEE 802.3 framing is not implemented on UWB, UWB loses much of it’s appeal. But if IEEE 802.3 framing is implemented on UWB, UWB might be MARKET ESTIMATES the future engine behind solutions we know as IEEE The Allied Business Intelligence research company has 802.11 based WLANs these days. estimated that 45 million UWB chips will be sold at 2007. It is also possible that we see two totally different market Worth of these shipments would be 1.4 billion euros. segments building up. These segments most likely will be First deliveries to buyers is estimated to happen in the one for WPANs, multimedia device networks and cable end of 2003, after which UWB will increasingly grow its replacement and one for WLANs. This set-up leads UWB market share of the home wireless market according to compete with Bluetooth and many of the wired alterna- In-Stat/MDR, the high-tech market research firm. Despite tives, USB and Firewire as most notable contenders. that UWB will be strong rival to IEEE 802.11 based Regulatory bodies like FCC, ETSI and ITU among oth- WLANs; existing WLAN standards will dominate mar- ers, play an important role in wide acceptance of UWB kets throughout 2006. because UWB is fundamentally different from traditional In-Stat/MDR further estimates that the first UWB prod- frequency oriented radio communications paradigm. Use ucts will be designed for the WPAN environment, and of UWB consumes huge amount of frequencies but in will have speeds around 100 Mbps, and a range of 10 m. very limited area and time span. The first UWB products will probably come from the In February 13, 2003, the US Federal Communications consumer electronic and set-top box companies. Other Commission (FCC) allowed use of UWB devices under UWB companies that are planning to enter the home net- FCC rules, part 15, and reaffirmed the regulations it passed working market include Time Domain and General Atom- on February 12, 2002 with some minor changes that eased ics. the design of UWB devices. Still, FCC reserved the right All these estimations comes with condition that the FCC for further regulations after 12-18 months period that will will allow full potential of UWB. Uncertainty inherited provide feedback and real life experiences. (FCC 2003) from FCC’s reservation of rights for future regulations Removal of regulatory uncertainty has major impact on may cause vendors to abandon this promising new tech- UWB market growth. While FCC has no rule setting nology. power on other countries, it has shown the development path that other regulators will follow. In fact, radio frequency licensing issues regarding UWB have fuelled the CONCLUSIONS reformation of FCC rulemaking process. (FCC/Powell 2003). UWB falls between IEEE 802.11 tecnologies and Bluetooth. IEEE 802.11 based WLANs offer greater coverage with narrower bandwidth for computer equipment while Bluetooth offers seamless communications for personal communication devices within smaller area but with slower speeds. Both of these two are existing and mature frameworks for short range communications. It seems that UWB is not direct replacement of these two REFERENCES technologies and there might be a market segment that technology companies may exploit (Kivisaari Eino, Tech- Articles nology Strategy Analysis for an Emerging Communica- tions Technology, pp 54.). In spite of that we believe that Timo Helenius, Buffalo Airstation WBR-G54, Tietokone, IEEE 802.11 technologies and Bluetooth dominate and January 2003, pp. 13. UWB will note replace it in shorter periods (before the Jarmo Huhtanen, Brice Clark: Kännykkäverkoissa väärää end of 2004). Customers do not want one more commu- tekniikkaa, Tietokone, November 2002, pp. 16. nication method just for technology reasons. Customers select one or two technologies and deploy it everywhere Samuli Kotilainen, Langattomat verkot rajussa it is possible whether it is optimal or not (Huhtanen 2002). myötäisessä, Tietokone, October 2002, pp. 10. Existing technologies have advantage on this kind of set- Petri Mähönen, Tulevaisuuden mobiilit dataverkot, up. This happens for simplicity’s sake and economical Tietokone, August 2000, pp. 62-70. reasons. Programmable radio transmitters and receivers may alter market situation and make communication hi- Jyrki Oraskari, 4G - Matkapuhelimien neljäs sukupolvi, erarchy more viable solution (Oraskari 2002). This may Tietokone, June 2002, pp. 103.

59 Studies White Papers / Specifications Jin Ding, Li Zhao, Sirisha R. Medidi and Krishna M. Federal Communications Commission, First Report and Sivalingam, “MAC Protocols for Ultra-Wide-Band Order Revision of Part 15 of the Commission’s Rules (UWB) Wireless Networks: Impact of Channel Acquisi- Regarding Ultra-Wideband Transmission Systems, April tion Time”, ITCOM 2002 (IT401), July 2002, http:// 2002, http://www.fcc.gov/oet/dockets/et98-153 www.eecs.wsu.edu/~jding1/paper/ITCOM401-02/ Federal Communications Commission, News Release itcom401-02.pdf. FCC Affirms Rules to Authorize the Deployment of Ul- J. Foerster, E. Green, S. Somayazulu, D. Leeper, Ultra- tra-Wideband Technology, February 2003, http:// Wideband Technology for Short- or Medium-Range hraunfoss.fcc.gov/edocs_public/attachmatch/DOC- Wirelss Communications, Intel Technology Journal, 2Q 231197A3.pdf 2001, http://intel.com/technology/itj/2001/articles/ Federal Communications Commission, Statement of art_4.htm. Chairman Michael K. Powell Re: Revision of Part 15 of Kivisaari Eino, 2002, Technology Strategy Analysis for the Commission’s Rules Regarding Ultra-Wideband an Emerging Communications Technology - Case: Ultra Transmission Systems, February 2003, http:// Wideband Communications, Master’s thesis, December hraunfoss.fcc.gov/edocs_public/attachmatch/DOC- 16 2002, Helsinki University of Technology. 231197A1.pdf.

60 60 GHz MMW Applications

EINO KIVISAARI TELECOMMUNICATIONS SOFTWARE AND MULTIMEDIA LABORATORY HELSINKI UNIVERSITY OF TECHNOLOGY P. O. BOX 5400, FIN-02015 HUT, ESPOO, FINLAND [email protected]

1 ABSTRACT WLAN networks or fixed wireless connections in metropolitan areas. 10 Modern multimedia applications demand higher data rates, and the trend towards wireless is evident, not only In the frequency range of 60 GHz systems (typically 58- in telephony but also in home and office networking and 64 GHz) there is no need for licensing in most countries, customer electronics. This has been recently proven by which makes the deployment of 60 GHz systems a lot the accelerating sales of IEEE 802.11 family WLAN hard- smoother operation compared to building e.g. 3G cellu- ware. lar networks. However, the 60 GHz radio technology is not a directly competing technology for 3G systems, but Current WLANs are, however, capable of delivering only instead a building block for future wireless networks with 10-50 Mb/s connection speeds, which is certainly insuffi- very much more bandwidth and also much less mobility cient if we look at the bandwidth requirements of future than current cellular systems. applications like wireless high-quality videoconferencing, multiple simultaneous wireless IEEE 1394 (Firewire) Generally speaking, the more speed we need the more connections or wireless LAN bridges across network seg- bandwidth we need. Transmission of several hundred ments. For these and many other purposes a lot more ca- megabits (or even a gigabit) per second requires very large pacity—wirelessly—is needed. It seems that today casual bandwidth, which is available in the millimeter wave area. Internet users and office workers are rather happy with Fortunately, there is a remarkably large frequency range the service level provided by IEEE 802.11 WLANs, but allocated for unlicensed wireless telecommunications the rising of users’ bandwidth demands does not show around 60 GHz. In Europe the frequency ranges 62-63 any signs of decline. GHz and 65-66 GHz are reserved for wideband mobile networks (MBS, Mobile Broadband System), whereas 59- What makes 60 GHz millimeter wave (MMW) systems a 62 GHz range is reserved for unlicensed wideband wire- very attractive solution for the purposes described ear- less local area networks. In the United States the frequency lier, is the fact that there is a several GHz wide frequency range 59-66 GHz is a generally unlicensed range. In Ja- range available around 60 GHz, almost worldwide. This pan 59-66 GHz is reserved for wireless communications massive spectral space (on the magnitude of 5 GHz) en- [3] [4]. In comparison to for example 17 GHz radio sys- ables densely situated, non-interfering wireless networks tems the 60 GHz has a huge advantage in terms of the to be used in the most bandwidth-hungry applications of spectral space available. the future, in all kinds of short-range (<1 km) wireless communication. In addition to the large unlicensed spectral space avail- Country / Region Frequency Purpose able, the propagation characteristics of 60 GHz radio sig- Range [GHz] nals are favorable for implementing very high speed wireless networks close to each other in metropolitan envi- Europe 62–63, 65–66 MBS ronments. Probably the most promising application areas 59–62 WLAN of 60 GHz systems are indoor WLAN solutions and fixed wireless connections, perhaps complemented with FSO USA 59–64 General purpose (Free Space Optical) lasers to improve outdoor reliabil- Japan 59–66 Wireless ity, where heavy rain can significantly attenuate 60 GHz Communication signals.

Table 1: Frequencies Available for 60 GHz Wireless 2 INTRODUCTION Communications Worldwide The 60 GHz millimeter wave (MMW) radio technology The data amounts transmitted in wired and wireless net- is a promising candidate for fulfilling the future needs for works have been increasing due to the ever growing popu- very high bandwidth wireless connections. It enables up lation of Internet users and the more and more feature- to gigabit-scale connection speeds to be used in indoor rich services used. This combined with the trends towards

61 wireless networking (convenience, flexibility) on one telecom operators) in a small geographical area, and net- hand, and increased need for wireless high-speed data work redundancy is easier to implement. transfer also in the consumer market (digitalization of consumer electronics such as digital still cameras or digital video cameras and video projectors) on the other, brings the need for new very high speed wireless technologies. As a technology per se, the 60 GHz MMW is not a new invention. As early as the 1960’s, several applications were implemented by military or governmental institutions but with immense costs.

3 SPECIAL FEATURES OF 60 GHZ MMW

Atmosphere attenuation In conjunction with the unlicensed frequency range benefits of 60 GHz systems, also the relatively high free space path loss and especially the extra attenuation caused by oxygen absorption make 60 GHz suitable for high density metropolitan area wireless networking. In these networks, the mutual interference of closely situated network islands (hotspots or wireless link lines) is dramatically reduced due to attenuation characteristics of 60 GHz systems. In the following graph (figure 2) is presented the signal attenuation (dB/km) as a function of signal fre- Figure 2: Signal Attenuation in Atmosphere as a quency. Attenuation caused by oxygen resonance is about Function of Frequency [2] 10–16 dB/km with 60 GHz signal (attenuation in excess of 10 dB/km occurs in a bandwidth of approximately 8 GHz, centered around 60 GHz). Rain Attenuation In addition, the free space path loss increases quadrati- cally with signal frequency, meaning that on 60 GHz fre- The most problematic condition for outdoor 60 GHz sys- quency the free space attenuation is 21 dB higher than on tems is heavy rain. 60 GHz radio signals are millimeter 5 GHz, for example. This might indicate an unfortunate waves that attenuate heavily when trying to travel through propagation behavior for 60 GHz compared to 5 GHz rain drops—the size of which is close to the signal wave- because of the 21 dB link budget disadvantage of 60 GHz length. This behavior limits the range of reliable (carrier systems. However, due to the lower co-channel noise of class) 60 GHz links to about 400 m (in average rain re- 60 GHz and the “cleaner” operating environment achieved gions). However, in indoor WLAN applications this is by heavy attenuation caused by concrete walls, 60 GHz not a problem—quite naturally—and in outdoor fixed systems can achieve similar or even better SNR (signal- wireless systems there are ways to substitute for rain out- to-noise-ratio) figures than 5 GHz systems. For example age. This can be done by using for example a HFR (Hy- an ordinary concrete wall attenuates 60 GHz signals ap- brid FSO Radio) system in which MMW transmitter is proximately 36 dB, which can be considered as isolation. used in conjunction with free space optical (FSO) laser This provides less noisy operating space for adjacent 60 beam system, which essentially unaffected by rain. Un- GHz cells and perhaps even more importantly a means of like rain, fog does not inhibit the propagation of 60 GHz information security [3]. An outstanding worry for cur- signals. And vice versa: fog attenuates FSO significantly, rent WLAN owners is the eavesdropping and resource but rain doesn’t—a fruitful symbiosis there! theft done by individuals with WLAN equipment just walking or sitting near an office building—still well within the WLAN network’s coverage. Indoor behavior This rather original behavior of 60 GHz radio signals (the Also in indoor WLAN applications 60 GHz systems have rapid attenuation in free air) is probably the main reason special features. The diffraction of millimeter waves is why the sharpest edge of 60 GHz research work today is weak, and this makes 60 GHz signals vulnerable for physi- in Japan, the country of extremely dense city infrastruc- cal obstacles such as human beings in an office environ- ture. In Japan, the capability of 60 GHz MMW systems ment. For this reason, omnidirectional user antennas are for very dense frequency reuse with any given area is of favorable as they can take advantage of reflected signals crucial importance. Many 60 GHz systems can be oper- in case of line-of-sight (LOS) obstruction. ated by several operators (firms, individuals or even Walls are a blocking obstacle, and for this reason every

62 indoor environment (room, hall corridor) needs to have at least one 60 GHz access point installed. For diffraction reasons the best possible mounting position would be in the middle of the ceiling, but systems are being designed so that MMW access points could be installed on walls— near existing LAN cablings—instead of ceilings.

4 MAIN APPLICATIONS FOR 60 GHZ

4.1 WLAN The emergence of 60 GHz WLAN networks in popular use requires mass production of small, inexpensive and highly integrated components. For this reason we haven’t seen 60 GHz systems in WLAN use yet: the unit prices for essential components have been going down but not yet reached a level suitable for mass production of consumer WLAN hardware. One point to ponder is the re- Figure 3: Multiple FiberLeap OC-12 Channels In a met- quired semiconductor technology. Current 60 GHz de- ropolitan area [2] vices (amplifiers, mixers) make use of a relatively expensive GaAs process, while cheaper alternatives are searched. 4.3.1 HFR (Hybrid FSO Radio) Fixed wireless systems in outdoor installations often suf- 4.2 Mobile Networks fer from availability problems: FSO (Free Space Optical) Basically due to the rapid attenuation of 60 GHz signals, laser links suffer in foggy conditions and, on the other this is not a technology for cellular systems with wide hand, for example 60 GHz radio links have difficulties in geographical coverage. However, we do not know what heavy rain. An US-based company AirFiber the future will bring and how far the consumers are will- (www.airfiber.com) has brought to market a product that ing to push the limits of current and successive wireless combines these two technologies. The weather conditions technologies—with their purses of course. problematic for FSO and 60 GHz, accordingly, do not In some papers 60 GHz MMW is presented as the 4G overlap and hence it is possible to achieve a link reliabil- communications technology, but a more feet-on-the- ity of 99,999 %, AirFiber claims. In the industry, this kind ground guess would place the 60 GHz in WLAN & Fixed of system is called HFR (Hybrid FSO/Radio). Currently Wireless applications. AirFiber states it can provide fixed wireless link systems with speeds up to 1.25 Gb/s, regardless the weather.

4.3 Fixed Wireless Along with the future potential in WLAN systems, fixed wireless seems to be the most promising application field for 60 GHz MMW technology. 60 GHz systems with less than 500mW output power can be widely used without licensing, which relieves the installer from the unrecurring but large initial costs associated with licensing. In addition, in many geographical areas only two or three carriers are allowed to operate on licensed spectrum areas [1]. Another issue is reliability and interference. Compared to for example UWB (Ultra Wideband) communications, 60 GHz MMW systems make use of a proven technology with little worries about interference. Up to date, UWB Figure 4: Predicted MMW and HFR maximum ranges has shown great potential as far as power consumption, in various rain regions for 99.99 % statistical path frequency reuse and manufacturing costs are considered— availability [1] (Scale maximum is 1600 m. The taller columns are 60 but still unresolved questions exist about interference and GHz+FSO, shorter ones 60 GHz only.) regulations. Fixed wireless applications of 60 GHz technology are commercially available today but it remains to be seen wheter UWB can ever be used for this pur- A point to take in fixed wireless considerations for 60 pose. GHz is that attenuation, the 10-16 dB/km regime, makes

63 long-distance connections (>2 km) virtually impossible. optimized. However, in the indoor applications mentioned before the The user hardware antennae for 60 GHz WLANs should 10-16 dB/km oxygen attenuation has no significance what- also be optimized for better behaviour when line-of-sight soever. is lost. All in all, the technical problems listed above seem to be of a nature that requires just hard work. The biggest question, however, is in the overall timing of 60 GHz com- mercialization. As noted before, fixed wireless solutions exist already (FiberLeap, AirFiber, even Nokia MetroHopper), but 60 GHz WLAN products are far from the retail store shelves. If we look at the speed with which the IEEE 802.11 WLAN generations have succeeded each other, we might come to a conclusion where it takes many several years before 60 GHz WLANs are anywhere near mass production.

6 CONCLUSIONS 140 All in all, 60 GHz is a frequency we all are certainly going to hear about in the future—if not for other reasons then because it is really the first time in history when a frequency space as huge as 5 GHz has been allocated for unlicensed communications use, set aside the 3 GHz worldwide overlap in these frequencies, also totally un- precedented. The evolution, or commercial roll-out, has started with fixed wireless metropolitan area solutions, and it is likely to continue with WLAN-like applications later on. Perhaps one of the most interesting questions is whether 60 GHz or UWB will dominate WLAN and home wireless markets after the leap to ultra high wireless speeds is made – perhaps during the next decade. UWB has significant advantages in power consumption, cost, and circuitry simplicity, but on the other hand 60 GHz has the huge reserved spectral space and more proven and more reliable technical background. Both may also be able to survive: UWB in home applications and customer electronics, 60 GHz in corporate use Figure 5: HFR (Hybrid FSO/Radio), a prototype with higher security and reliability requirements. Many system by AirFiber things depend on what kind of resolution is found for UWB regulations worldwide—but even regardless of that, 60 GHz systems will eventually make their way to wideband wireless market. 5 GHz of specifically allocated spectrum 5 PROBLEMS, FUTURE WORK is surely not going to be abandoned useless. 128 Despite the many good news about 60 GHz there are * * * still many problems to be solved before mass production roll-out. In designing systems and protocols for 60 GHz MMW communications, the main guidelines for architecture de- Perhaps the most decisive and also very straightforward sign should include affordability, scalability, modularity, issue is the lack of cheap enough, miniaturized modules extendibility and interoperability [3]. For 60 GHz con- for hardware implementation (integrated amplifiers, mix- sumer mass markets to take off, it is crucially important ers etc. and antennae). Current GaAs semiconductor plat- to design such architectures that enable product differen- forms are likely to be replaced by cheaper technology, tiation and modification for many different usage sce- but the industry is not there yet. narios. One key aspect is the selection of suitable modu- One attractive issue to work on is the integration of 60 lation scheme to foster 60 GHz equipment industry, help- GHz radio frequency circuitry with the antenna so that ing it to make various gadgets optimized for mobile use— intraconnectivity losses could be reduced and size/costs all being compatible. OFDM seems to be to the solution

64 for that, but it is another story. April 7, 2003) [3] Smulders P: Exploiting the 60 GHz Band for Local Wireless Multimedia Access: Prospects and Future Di- 7 REFERENCES rections, IEEE Communications Magazine, pp. 140-147, [1] Bloom, Hartley: The Last-Mile Solution: Hybrid FSO January 2002. Radio, AirFiber, Inc. May 2002, http://www.airfiber.com/ [4] Mikkonen J: Emerging Wireless Broadband Networks, products/HFR.pdf, (Referenced April 8, 2003) IEEE Communications Magazine, pp. 112-117, February [2] Telaxis Communications: 60 GHz, the optimum fre- 1998. quency for wireless metropolitan networks, www.tlxs.com/ products/pdf/why60.pdf, www.fiberleap.com (Referenced

65 Introduction to Software Defined Radio

KAI KUIKKANIEMI INSTITUTE OF STRATEGY AND INTERNATIONAL BUSINESS HELSINKI UNIVERSITY OF TECHNOLOGY P.O. BOX 9500, 02015 HUT, ESPOO, FINLAND [email protected]

ABSTRACT The road that led to development of SDR was affected by several different trends. The first and the most important Software Defined Radio (SDR) is not, like normally in enabler of SDR is naturally new hardware. If the hard- technical terms, referring to any single technology, but it ware wouldn’t be that flexible and suitable to be pro- means the idea of controlling radio functionality (for ex- grammed via software, there would be no point in talking ample operating frequencies and modulations) both in about SDR. SDR doesn’t mean that hardware is replaced; transmitter and receiver side by manipulating flexible actually quite vice versa it sets rather high demands for hardware components with software. SDR functionality the hardware. can be implemented in several different levels so that it covers from one component potentially all relevant radio When we are talking about hardware there are two funda- parameters. There are several different applications for mental, sort of abstract, parameters that should be kept SDR and in addition there are also different architectures separate. First of all components must have a wide func- and methods how it can be done. tional area. Another important feature is that the hardware is programmable and has proper interfaces. In addi- SDR has definite uses in future when the wireless net- tion to hardware there are some software requirements working environment becomes more and more complex. for the SDR. But at the same time it could provide benefits there are also few technical and regulatory problems involving to SDR enables sort of ideological thinking, meaning that SDR, which must be first cleared out. with SDR there is freedom and flexibility of new applications. If somebody has SDR enabled base stations and terminals they can practically invent their own communi- 1 INTRODUCTION cation standard, which is a great possibility in terms of innovations and understanding. At the same time there is Origins of Software Defined Radio development can be of course some problems relating to possibility to hack found from military [1], where its flexibility was discov- existing networks. The requirement for easily configurable ered to be very useful quality already in early nineties. and manageable network structures comes more and more Still currently the largest and most significant implemen- important when we start to think SDR enabled AdHoc tations of SDR are related to US military. and P2P networks. Current radio architecture (even the digital one) basically Last but not least the important driving force of SDR has dates back to 70’s. Meaning that for a long time the radio been the development of common, open and non-propri- technology is based on similar type of rigid component etary standards and protocols. If the standards and proto- structure. SDR could be the first disruptive technology in cols where closed, it would be much less interesting to this area [2]. Currently SDRForum is responsible body have easily adaptable devices. Of course military and even orchestrating SDR related activities. some business type application of SDR make sense al- All the current technological transitions (digitalization, ready without open protocols, but the biggest possibility packet data, broadband and service quality) in ITC in- of SDR to spread around and liberate common person dustry can be seen to support the SDR based technology communication requires open protocols to tap in. ramp-up [3]. These transitions clearly indicate that we are evolving to a situation where we have high performance networks of networks. In addition to high perfor- 2 FUNCTIONAL CHAIN OF RADIO (AND mance it is also important that the communication can HOW IT APPLIES TO SDR) take place anytime, anyplace and in combination of sev- The following structure I explain here is more or less the eral modes like voice, text, data, and video. Besides the same as in all common digital wireless radio networks modes there are also different communication formats, structure from transmitter to receiver. But still understand- push, pull and multicast, just to name the most common. ing and revising this is very important to understand where This kind of network environment clearly requires also and how SDR has effect on and what is SDR’s playground. flexibility on radio components, which is something that The following is based on Techtargets [5] definition and SDR can provide.

66 bit altering from normal it is tailored as a SDR definition. 20-500MHz, 1-2GHz) under software control without hardware change. These systems are typically character- A typical voice SDR transmitter, such as might be used ized by a separate antenna system followed by some in mobile two-way radio or cellular telephone communi- wideband filtering, amplification, and down-conversion cation, consists of the following stages? Those compo- prior to receive analog-to-digital conversion. The trans- nents that could be tailored and controlled with the soft- mission chain provides the reverse function of direct digi- ware are marked with an asterisk (*). tal-to-analog conversion, analog up-conversion, filtering, 1. Microphone and Audio amplifier and amplification. 2. Analog-to-digital converter (ADC)* Tier 3. Ideal Software Defined Radio 3. Modulator * This system has all of the capabilities of the Tier 2 sys- 4. Amplifiers tem, but eliminates analog amplification or heterodyne 5. Transmitting antenna * mixing prior to digital-analog conversion. It provides dra- A typical receiver designed to intercept the above-de- matically improved performance by eliminating analog scribed voice SDR signal would employ essentially the sources of distortion and noise. same stages reversing the transmitters order and function Tier 4. Ultimate Software Radio (hence amplifiers are replaced by superheterodyne sys- This would be the dream situation where the SDR based tems, modulators by demodulators and ADC’s by DAC’s). device is small, fully configurable, ideal in terms of all This basically shows that SDR can play effect practically different protocols, quick to respond and contains good to most of the hardware components in radio. Picture be- processing power to utilize additional network services. low further illustrates in more detail components inside This tier description is intended for comparison purposes radio that SDR plays role (picture copyright SDRForum). rather than implementation.

4 TECHNOLOGY

Antennas Software radio design begins at the antenna. In order to fully utilize the software radio approach it is important that antenna has good beam forming, diversity and sectorization qualities [7], [8]. Wideband (WB) and ultra wideband (UWB) antennas are antenna technologies, which enable accessibility to multiple RF bands dynamically, sequentially or in parallel. These wideband antenna 3 TIERS technologies are aligned with the development of micro electro-mechanical systems (MEMS), which can replace Already in the first chapter it was mentioned how the SDR many different components like pin diodes. Advances in as a phenomenon can take place in several different tiers. the simulation techniques enable better management of More accurate tier definitions can be found from these antenna technologies, when their behaviour in dif- SDRForum website and it basically reflects pretty well ferent situation can be modelled before hand and the sig- how they view the SDR’s possibilities and development nal tailored to the antenna qualities. [6]. Tier 0. Hardware Radio The tier 0 is not basically SDR at any ways, but is the Hardware enablers [8] most common current situation of the radio devices. No Basically also antenna should belong to this category, but provision is made for any changes of system attributes I wanted to emphasize the importance of the antenna as except by physical intervention by the user or a service an enabler of well functioning SDR, so I made own chap- technician. System operation is accomplished by use of ter for it. The hardware enablers can be roughly divided switches, dials, and buttons, by physically opening the in to two chapters RF enablers (which contains basically covers, or by replacing the unit. RF circuits, antennas and MEMS) and to digital enablers Tier 1. Software Controlled Radio (processing devices and communication fibres). Tier 1 is the most elemental level of SDR. Radios in this Due to very wide operating frequency (possibly 2Mhz – category have control functionality implemented in soft- 5Ghz) very specialized RF technologies are needed for ware, but do not have the ability to change attributes, such SDR. Multiband (MB) and multimode (MM) chipsets are as modulation and frequency band without changing hard- current answer to the requirements of the SDR. MEMS ware. are the components that are backbone enabling the order Tier 2. Software Defined Radio of magnitude improvements in MM/MB chipsets. Fur- The Tier 2 system provides a broad operational range (e.g. ther improvements are expected from novel RF signal processing methods.

67 MEMS can be used in various points (like explained al- In the commercial ITC space the most anticipated prod- ready earlier) in these new chipsets. Examples of compo- uct is naturally software phone which could potentially nents that MEMS can replace are bulky pin diodes, su- support different cellular networks (2G to 3G and from per-wide field-effect transistors (FET) and vacuum tube GSM to CDMA), different wireless LAN network, per- relays (VTR) in antennas. They can also be used as high- sonal area networks (like Bluetooth), UWB networks and performance miniature inductors, capacitors, filters, T/R GPS signals with the same radio chipset, while at the same switches and diplexers in RF front ends. time it would have potential to adapt new yet unknown standards. Advantages are obvious, but there are still sig- Advanced Field Programmable Gate Arrays (FPGA) prod- nificant problems in the area due the increased power uct families are providing “system-on-chip” feature. These consumption and complexity. technologies basically contain embedded serial transceivers, RISC processor and some amount of programmable Amateur radio [2] is perhaps the area where the biggest memory. In the end they provide the digital front-end for interest for open SDR system origins. The community of software configurable radio signal processing. Also CPU radio amateur worldwide could start to use more the com- should and will [9] be supporting SDR. puter as an interface for their communication. This community is experimenting all the time different models of and protocols of communication, and they would just love Software enablers [8] if they would have an open platform to tailor the chipset Software enablers of SDR can be divided in to four cat- functionality to different setups. When they are currently egories. Software Communication Architecture (SCA), heavily limited to the features that their hardwired receiv- Waveform Development Environment (WDE), Radio ers and transmitters can provide. Description Language (RDL) and Radio Virtual Machines. At the same time the open platform with the radio ama- In addition to this there is need for real time operating teurs could really liberate the communication environment, system. and potentially generate many new interesting innovations Software Communication Architecture (SCA) means ba- when a large pool of people have access to radio tech- sically the platform on which SDR can be build. Most nologies, it contains also some problems. How to prevent promising development in this area is run by Object Man- radio amateurs (radio hackers) from disrupting licensed agement Groups (OMG) SDR special interest group. Their traffic. platform independent model (PIM) will be mapped in In addition to mentioned case examples military case minimum version to for example J2ME and in server side SpeakEASY will be presented later on. Then I left out to for example CORBA and .NET environments. several civil applications and base station applications, Most of the WDE toolsets employ “mainstream” simula- which are yet also interesting, but at this point one should tion environments such as MATLAB and SIMULINK. be able to already grasp the concrete benefits SDR can Because of the platform specific optimization of IP cores provide. and DSP algorithms, under some circumstances SIMULINK generated code on DSPs and FPGAs may perform better than when hand-coded. 6 BUSINESS RDL is a higher order language originally used to “construct” a radio functional model using RDL “building General business benefits blocks”. RDL is used to configure a flexible modem, describe the desired signal processing graph, and give pa- When the hardware becomes more recyclable there is high rameters for each processing stage. It doesn’t include potential for significant life cycle cost reductions. Over implementations of signal processing stages, nor is it a the air downloads of new features and services as well as waveform specification language. software patches enable that the both network and terminals can quickly adopt and take advantage of new tech- The RVM is a hardware abstraction, which can signifi- nologies [10]. Advanced networking capabilities would cantly accelerate time to market. It brokers parallelism in finally allow truly “portable” networks. multi-core, multiprocessor, and accelerated designs. It allows the interoperability of multivendor, real-time in- In the end the frequency space we have is very limited, tellectual property at both ‘whole-stack’ level and ‘stack- and already now there are clear cases where the frequency component’ model. band has been exhausted. SDR could enable much more efficient usage of frequency band by quickly adapting to utilize all available free spectrum. This is further discussed 5 APPLICATIONS in with secondary bandwidth utilization later on. A com- prehensive list of value propositions for different players in the field (consumers, operators and manufacturers) can There are several different application areas for SDR rang- be found from SDRForum website [10]. ing from handsets to base stations and from military applications to amateur radio.

68 Cost – Benefit analysis can adapt to these network conditions intelligently without user intervention in order to maximize the wireless Following cost-benefit analysis is based on Joseph Mitola usability. At the same time different mobile network [11] analysis on the topic. One must remember that life components (mainly terminals) could form ad-hoc new cycle costs include R&D, acquisition and operations and networks to satisfy some temporary communication needs. maintenance components. Key techniques in accurate cost/ benefit analysis include parameterizing the relationship between the technical features of the system architecture, Secondary markets [3] the technical architecture and the projected revenue streams. Secondary bandwidth markets is interesting new concept in order to business vice utilize the network capacity at But still in the end in case of SDR everything comes down its maximum. Basically it means that primary buyers pur- to chip cost. Normally a dual-purpose chip cost (e.g. dual chase the operating licence for the frequency bandwidth, band GSM) around 1.5-1.7 times the cost of single pur- but in the secondary markets they could resell parts (that pose chip [11]. Then when there are more than two modes they don’t need or other parties can provide better value and or frequency areas that chip needs to function the added) of the bandwidth to other players. cost increases around again with respective incremental amount. Basically this would require some kind of network bro- ker, which would manage the secondary markets and sell Another factor that is very important especially for smaller the frequency space-time of the primary operator to sec- vendors and instances that would like to start implement- ondary operators. Good example where such secondary ing SDR products is the availability of COTS components markets would be very beneficial would be for example for SDR. irregular events, like Olympics, that have an exceptional need for all different bandwidths. For these secondary markets there is need to have easily configurable termi- Market overview nals and other network components in order to maximize Worldwide interest and investment in the SDR technolo- the usage of continuously changing network conditions. gies is growing significantly, with key standardization and development efforts now taking place throughout Europe, North America, Japan, Korea, and China. In addition to 7 CURRENT SITUATION the broad benefits listed in chapters above, SDR technologies offer unique benefits to players on every tier of the value chain [10]. Case examples Governing bodies like FCC (discussed in next chapter) SPEAKEasy [13] is the US military development project are already heavily analysing SDR products and in SDR of SDR, that started already in mid 90’s and is now al- Forum there are currently 122 member companies and ready in second phase of implementation. The key vision organisation (not Nokia, just to point out in Finnish per- in SPEAKEasy development was to build the “PC of the spective), which contribute to the forum activities mon- Communications World”. This basically means a fully etarily and otherwise. programmable waveform and COMSEC for Voice, Mul- timedia and Networking Use. It is designed to be multi- Most of the companies that are developing already actual band system operating continuously. It has both open products are doing that for military or are technological modular hardware and software architecture, which has spin-offs of military research for commercial purposes. been a basis for first commercial installations also. There are few cases explained in later on in the chapter. Communications Research Centre Canada (CRC) and Defence Research & Development Canada (DRDC) were Future prospects the first ones to show a commercial demo of SDR radio The market penetration of SDR technologies is analysed environment [14]. This happened in November 2002. In to be concurrent with the ramp-up of 3G networks[12]. addition there are few companies like Radical horizon Cognitive radio [13] is then more in future located vision [15] and SDRT [16] which have near ready implementa- of what SDR utilization might look like. The idea in cog- tion of SDR for cellular networks. nitive radio is that all radio equipments would sense all In parallel to SDRForum’s effort to harmonize SDR de- the time surroundings and communicate with each other velopment there are some other “open” movements to in order to create the most efficient and robust network- build SDR like system. Good examples of these are GNU ing conditions. For example pacemaker could have alter- open radio [17], which is mainly a amateur radio based native operating frequencies, which could be immediately initiative and Open radio platform [18], which tries to taken in to use if the primary operating frequency is occu- build a 3G environment for educational and research pur- pied otherwise. poses. Basically SDR is a logical step in intelligent radio networks where the base stations provide different capaci- ties and functionalities in different places and terminals

69 Problems Mitola When SDR systems becomes widely available to small 5. Software-defined radio, http://searchnetworking. developer communities like previously introduced radio techtarget.com/sDefinition/0,,sid7_gci333184,00.html ,7/2001 amateurs, there is a problem of them to accessing exist- 6. Definitions & Semantics, http://www.sdrforum.org/ ing networks and disrupting licensed network traffic. tech_comm/definitions.html, There are some ideas how to prevent this from happen- 7. Seven steps in designing software defined radio, http:// ing, like restrictions and limitations in the software, but ourworld.compuserve.com/homepages/jmitola/seven.htm, Jo- so far there is no consensus or clear vision how this should seph Mitola be done. 8. Research and development working group 2002 summary report http://www.sdrforum.org/public/approved/03_a_ But even tough clear regulatory problems, FCC [3] states 0002_v0_00_rd_sum_01_30_03.pdf , David Murotake, 1/2003 that they are “very interested in the area and look forward to its potential application”. Especially in the perspective 9. Intel CTO at Intel Developer Forum, Pat Gelsinger, 2/2002, of secondary markets the governing bodies see that SDR 10. Value proposition of SDR, http://www.sdrforum.org/ may become very beneficial. mrkts_comm/value.html, The biggest technical problem of SDR is high power con- 11. Cost-Benefits analysis, http://ourworld.compuserve.com/ sumption of hardware components. homepages/jmitola/costbene.htm, Joseph Mitola 12. SDR and wireless network adaptation, http://www. sdrforum.org/mrkts_comm/adoption.html 8 CONCLUSION 13. SPEAKEasy Military SDR, http://www.its.bldrdoc.gov/ Software Defined Radio (SDR) is an idea of changing meetings/art/art98/slides98/bons/bons_s.pdf ,Wayne Bonser, 7/ hard coded radio circuits in to software programmable 1998 ones. In a future complex networking environment such 14. Press release, http://www.crc.ca/en/html/crc/home/ approach contains several benefits like the flexibility of mediadesk/sca_demo , Claude Blelisle, 1/2003 introducing new standards, maximizing the bandwidth 15. Company website, http://www.radicalhorizon.com utilization and opening the network development for bigger pool of vendors and developers. 16. Company website, http://www.sdrct.com 17. Initiative website, http://www.gnu.org/software/gnuradio/ Openness is one key fundamental of SDR initiative, which gnuradio.html basically means that more and more instances would have access to understand, utilize and further develop network 18. Initiative website, http://www.wireless3g4free.com/ technologies. SDR is not limited to any specific technology, but the solution areas vary from amateur radio and military application to a cellular networks and short- range wireless networks. SDR is an idea that can materialize in many different levels depending on which components used in the RF-cir- Abbreviations cuits and modulation processors support software controlling. 2G Second generation mobile phone networks e.g. GSM 3G Third generation mobile phone networks e.g. UMTS ADC Analogue to Digital Converter REFERENCES ASCII American Standard Code for Information Interchange The references are mostly web pages, which causes that COMSEC COMmon wealth SECurity many times the exact author and/or date is missing. So COTS Commercial Of The Shelf according information is provided in following if available. DAC Digital to Analogue Converter DSP Digital Signal Processor FCC Federal Communications Commission 1. Software defined radio, http://www.arrl.org/tis/info/ sdr.html, American Radio Relay League, 4/2003 FPGA Field Programmable Gate Arrays 2. Understanding Software defined radio, http:// GNU Recursive acronym Gnu is Not Unix www.smartmobs.com/archives/000339.html , Howard HDTV High Definition TV Rheingold,11/2002 MEMS Micro Electro-Mechanical Systems 3. Trends in the field of networks Software Defined Radio: A RF Radio Frequency Regulator’s Perspective http://www.fcc.gov/oet/speeches/ sdrforumsph.html , Dale N. Hatfield (FCC), 6/2000 SDR Software Defined Radio 4. Congnitive radio motivation, http://ourworld. UWB Ultra Wide Band compuserve.com/homepages/jmitola/cognitiv.htm , Joseph WB Wide Band

70 MPEG-4 Technology and Business Strategies

SONJA KANGAS MIHAI BURLACU HELSINKI UNIVERSITY OF TECHNOLOGY HELSINKI UNIVERSITY OF TECHNOLOGY [email protected] [email protected]

ABSTRACT content. It also brings multimedia to new networks and devices, including employing relatively low bit rate and mobile ones. Low-bit rate multimedia streaming over IP- networks is the next natural step from the market after the MPEG-2 MPEG-4 enables the production of reusable flexible con- success with DVD and satellite digital broadcasting. Me- tent. The issue of IPR (Intellectual property rights) has dia encoding standard MPEG-4 is a key tool when creat- been more carefully taken under consideration compared ing rich interactive media for wireless terminals. Through to previous standards. For network service providers wireless communication and media on-demand systems, MPEG-4 offers transparent information, which can be MPEG-4 has become very engrossing standard. The fu- interpreted and translated into the appropriate native sig- ture prospects are bright for MPEG-4. In this paper we naling messages of each network with the help of relevant shall present the technical description of MPEG-4 as well standard bodies. It uses improved concepts for visual and as the business strategies accompanied with a case study. audio to handle both natural and synthetic nature of the objects: media objects. Description for the composition Keywords: MPEG-4, encoding, video, mobile multime- of media objects has many new concepts with respect to dia, codec. the old versions of MPEG. Multiplexing, the process where separate parts of the video streams are joined together into one file, and synchronization for the data as- INTRODUCTION sociated with media objects is available for transport over network channels providing QoS appropriate for the na- Rich multimedia including visual communication and ture of the specific media objects. Interaction with the video capabilities is one of the key development areas in audiovisual scene generated at the receiver’s end is pos- the digital mobile media. The anywhere-anytime idea sup- sible. ports the advancement of mobile multimedia delivery. The idea of “anywhere” implies that the technology must be portable, and operate under low power mobile conditions: on a train, in the airport, in your car, on the move. “Any- TECHNOLOGY time” means that the consumer will decide when to use Moving Picture Experts Group (MPEG) have developed this technology, when to communicate, or when to access MPEG-4. MPEG-4 belongs to MPEG-standard family that information, according to his or her own schedule. These contains five media compression standards: MPEG-1, demands necessitate capable encoding standards and tech- MPEG-2, MPEG-4, MPEG-7 and MPEG-21, used for nologies for rich multimedia content, as in the future the different purposes, digital television, video conferencing, wireless bandwidth will be valuable and delivery of the audio compression and DVD-technology, just to name a best audio and image quality at the lowest bandwidth will few areas. MPEG-4 was released in 1998 and another be a key part of developing services around wireless digi- version in 2000 (newest version MPEG-4 AVC at the end tal media. Besides the usage in mobile area, enhanced of 2002). In MPEG-4 the audiovisual presentation con- video compression technologies will also play a central sist of separate and independent media objects that can role in the development of digital television, real time be e.g. sound, video, still image and 3D objects. interactive applications, synthetic content creation and animated chats just to name a few areas of implementation. MPEG-4 also plays a central role in the application In short, MPEG-4 is a set of specifications for: development platform; it is expected to be a core tool for the applications in 3G-handsets and other video-capable wireless devices. 1) Represent aural, visual and audiovisual content called media objects. These objects can be natural or synthetic. MPEG-4 standard is build on the needs of authors, service providers and users to get higher flexibility and to take more advantage of the fast deploying internet. For 2) Deliver these independent layered media objects over end users MPEG-4 brings higher level of interaction with heterogeneous networks by using streaming protocols (composition of media objects).

71 3) Render and present scenes dynamically (compared to the web).

4) Make it possible for the receiver to interact with the audiovisual scene generated at the decoder phase.

MPEG-4 is notably enhanced compared to MPEG-1 and MPEG-2. MPEG-2 and MPEG-1 were targeted mainly on natural content transmission (digitally television, video disks, DVDs). Dealing with synthetic content was a natural step to be supported in MPEG-4. Synthetic content handling was inspired by Virtual Reality Modeling Lan- guage. The idea of compression has been shifted into object-oriented compression of image and sound. They are Image 1: MPEG-4 profiles used under the generic term of “media objects”. The ob- As the media encoding/decoding is the main task of jects in the image (table, dog, hat) are identified and MPEG-4 standard, video codecs are naturally the key handled separately. Audiovisual objects (often named as building blocks for a host of new multimedia applications media objects in literature) types can be: natural video, and services for streaming video on the Internet, digital sound, compressed speech, synthesized speech, poly- television or wireless solutions. Codec stands for Coder/ phonic music (containing MIDI etc.), 2-D, 3-D objects, Decoder. It is a piece of software or a driver that adds a animated 3-D faces which can be synchronized with the support for certain video/audio format for the current synthesized speech [1]. operating system. With codec, the system recognizes the The standard is organised in a layered model. Briefly, most format the codec is built for and allows to play the audio/ important layers are the delivery layer, the sync layer and video file (=decode) or in some cases, to change another the multimedia layer (also named as compression layer). audio/video file into that format (=(en)code) [2]. Delivery Layer is used to transport the media objects in a streaming manner. DMIF-application interface (DAI) is used as an interface to the Sync Layer (used for synchro- BUSINESS STRATEGIES nization of elementary streams). Multimedia layer is the core of MPEG-4 standard. The concept of descriptor is MPEG-4 standard is in a continuum of video encoding used to specify the type of information identified in each MPEG-standards for several different purposes but also stream (descriptors identify group of elementary streams more closely concentrated on delivering rich multimedia to one media object). Scene description information is to mobile terminals. The role MPEG-4 shall have in the needed to define the spatial and temporal position of the market in the future in the “low-data rate” niche. Its media objects and their behavior over time. BIFS (BInary strengthness is that it is the only valid choice at the mo- Format for Scenes) is used to describe scene composition ment considering the air interface supported rates, and information. the processing power of the mobile terminal devices. Natural audio-video coding algorithms are based on block Although MPEG-4 is not as widely used as MP3 or the based DCT-coding, quantization, run length and huffman DVD, it already has started appearing in consumer prod- coding. Shape coding, dynamic natural image texture cod- ucts in the guise of QuickTime 6 and RealPlayer 8. With ing, static textures, error resilience (with the important MPEG-4, new opportunities will open for service pro- usage of reversible variable length codes) are just few viders. They can deliver new class of programming for tools used in MPEG-4 to provide the desired transmis- accelerating business areas including networked gaming, sion parameters. eLearning and media on-demand. The extensiveness of the MPEG-4 standard, it allows different players to spe- The basic requirement of having variable media bit streams cialize in their area of excellence. This will also bring within a wide range of bandwidth led to the approach of new possibilities for innovation processes and service defining the video objects in a layered manner. Each layer development when interlinking different experts together. has its own resolution properties. For the wide variety of usage, specific profiles and levels have been stated in the The growth of MPEG-4 is based on the market situation standard. Profiles are basically are group of tools that exist where the mobile reception is increasingly important and at a number of levels. The profiles define how and what rich multimedia is the next development step in mobile tools should be used for creating the bit stream, the levels markets. San-Jing Park, the CEO of Samsung, stated at define what values every parameter can have in a [* Filter the 3GSM World Congress 2003 that the consumer wants does not support this file format | In-line.TIF *]particular “anywhere, anytime, anything” [3]. In mobile industry, profile. the market of mobile entertainment, MMS services, mobile music and Java games is increasing. The values making it possible are accessibility, interoperability and

72 interactivity and the development of streaming video and videoconference applications can not compete with the rich media is enabled by standards like MPEG-4. fixed infrastructure networks. It is clear that the standard will co-exist with the other standards, without replacing There are no clear obstacles for MPEG-4 but several com- them completely. The investments recover for previous petitors like Microsoft (WindowsMedia) and Real standards like MPEG-2 prevents the immediate adoption. (RealPlayer), that might make the business situation more difficult. The most fundamental issue is the rate of change of MPEG-4 standard. Already it is very wide standard and as the user expectations and the media technology MPEG-4 BUSINESS AND A CASE STUDY advances, also the rich media compression standards MPEG-4 business can be analyzed as a standalone prod- should develop at the same pace. MPEG-4 has been uct, characterized through the success of products that claimed to be too large and slow in its changes. Despite incorporates it, and on the level of global acceptance in of its comprehensiveness, the slow reaction speed might the development industry market and end-users. It is hard invalidate its values if it is not up-to-date, and give con- to separate the development industry from the consumer. siderable advance to its competitors. Otherwise MPEG-4 There are three main technologies that will enable video is supported well and major players of interactive multi- on mobile devices in the near future: media are behind the standard. In wireless world, the main values of MPEG-4 are the wideness of the standard as it 1) The growth of data bandwidth in mobile networks covers different media formats and multi-player possi- 2) Improved video compression technologies bilities in wireless environment, and the most notable strength of the standard is in low-bandwidth rich media 3) Improved camera and display technologies [4]. delivery. Currently video telephony and conferencing, MMS and video messaging and video streaming have been used in mobile video applications. In this paper we present the One important drawback of the wireless systems is that DivXNetworks as an example of MPEG-4 developer. the computing power of the mobile devices is rather low. DivXNetworks is a strong market player. It offers a broad This means that high quality transmissions still require line of products for the delivery of DVD-quality video performant and expensive terminals. Also, in the moment over Internet protocol networks. Provides solutions for a of writing this report, the wireless networks does not of- wide variety of markets and businesses seeking to use the fer the data rates needed for high quality transmissions, digital delivery of video, the robust and complete prod- especially when long distances are involved. Also Qual- uct set is based on the MPEG-4 compliant DivX video ity of Service is not perfectly operational. Under these compression technology. DivX™is the main codec circumstances the videoconference application will be for (player) developed by DivX. DivX Pro™is a video com- a while still restricted to small data rates. The natural sup- pression technology designed specifically for professional ports for low data rates makes MPEG-4 the ideal choice users. to be used. DivX™for Gamecube Software Development Kit (SDK) The MPEG-4 standard is build on the needs of authors, is targeted to developers, providing video compression service providers and end users to exchange data at high technology within Gamecube games. Supported rates start- quality with significant low rates. This wide range of scope ing from 20 Kbps are naturally supported. Possible rates makes it more difficult to provide overlying applications. of 4000 Kbps on the Gamecube for a maximum resolu- The technology developers have a strong faith on MPEG- tion for decoding of 640x480 pixels. 4 standard as it has been stated that e.g. the cable industry intent to replace the current MPEG-2 implementations Open Video System™(OVS) is available for distribution with MPEG-4 in the very near future. of video on demand over broadband IP networks. The system was extended to OVS Network that uses a three- For authors the value in MPEG-4 is tied into the evolu- tier architecture and has been designed for digital video tion of integrated communication and multimedia devices distribution over broadband IP networks. like telephone, hi-fi, Internet, PC, television and video recorder all implemented in a single device. The full ad- The DivXNetworks is just one player at the market but a vantages of such integration will be realized when a com- key example on wide implementation of MPEG-4 stan- mon multimedia data format like MPEG-4 is used. dard. The other companies are following the same pattern, but the offer ranger is rather similar with the upper For fixed infrastructure networks, service providers will cases. For example Envivio has its own products and Real choose MPEG-4 especially for streaming applications due Networks provide the streaming server under the name to its better compensation performance with respect to “Hellix”. older MPEG-standards. High quality interactive applications (like videoconferences) are already reality. End users are expected to use Video-on-Demand (VoD) CONCLUSIONS applications in wireless area, since MPEG-4 is rather slow on encoding and fast on decoding. For the moment the MPEG-4 is wide and very strong standard. Its special quality of transmission for mobile terminals on characteristics include the presenting of generalized scene description language (BIFS), flexible composition of 2D

73 and 3D scenes and integration of multiple natural and and that it will be adopted as the preferred tool, espe- synthetic audiovisual data. It supports algorithms, tools cially in the IP streaming environments. Better compres- and profiles for an efficient coding of different kinds of sion and quality compared with other standard (consider- video, sound and speech data. MPEG-4 also allows user ing them at a particular transfer rate) switch will guaran- interactions with audio, video and graphics objects and tee its success. Especially in videoconference applications supports streaming of real-time data. As stated before, latency is a critical issue when considering transferring MPEG-4 operates on the levels of the authors, service video image. providers and the end users. Considering the MPEG-family for videoconference ap- It is widely believed that MPEG-4 type of formats will plications, MPEG-4 is a suitable candidate to be used, build new standards for multimedia in the future. Except because MPEG-1 is not generally used for applications audio and video also computer generated 3D images and where latency is an issue and MPEG-2 does not perform sound, still images, text and speech will be more widely significantly better than MPEG1, MPEG-4 with its new used together with the possibility of user interaction real streaming capabilities and QoS support can be a suitable time. It seems that www and multimedia formats will ap- option. proach each other. Www protocol already contains VRML, SMILE and interactive scripting languages like JavaScript and Flash. In the future the amount of interac- REFERENCES tive multimedia and video shall increase and it is highly presumable that MPEG-4 will have a central role in the [1] Koenen, R: Overview of the MPEG-4 Standard. http:// mpeg.telecomitalialab.com/standards/mpeg-4/mpeg-4.htm development. [2] WhatIs.Com (2003) Codec. Although MPEG-4 was set as a standard years ago, still- http://searchnetworking.techtarget.com/sDefinition/ unresolved licensing negotiations have held the technol- 0,,sid7_gci211810,00.html ogy back, leading to criticisms that its core technology, including its video formats, are out of date. Microsoft has [3] San-Jing Park, Key Note Speech at 3GSM 2003, Cannes France 15.2.2003. consistently led that argument in refusing to endorse a standards-based approach in the development of its Win- [4] Hantro (2002) White Paper: Video Codec Selection for Wire- dows Media multimedia technology. Other issues consider less Multimedia Terminals. http://www.hantro.com/en/ the evolution of wireless network technologies, the adop- whitepapers tion of new value web ideas [5] Dapend Wu et. al . (2000) On End-to-End Architecture for Transporting MPEG-4 Video Over the Internet. IEEE As many key players like Microsoft, Intel, Real Networks Transactionls on Circuits and Systems for Video Technology. and Apple have been contributed the development of Vol 10, No 6. MPEG-4, it is likely that the standard will have success,

74 Changing MVAS Environment

KIMMO PALLETVUORI HELSINKI UNIVERSITY OF TECHNOLOGY [email protected]

INTRODUCTION THEORETICAL BACKGROUND There are two major technical changes taking place in the area of MVAS. First, network technologies are changing Value Chain from traditional circuit switched cellular networks to more open shared access (packet based) networks like GPRS The traditional MVAS value change [mob02] is presented and UMTS. Second, mobile phones are changing to ver- in Figure 1. Portal company (MVAS provider) and Op- satile communication terminals accompanied with colour erator are both positioned to the right hand side of the displays and significant processing power allowing local value chain. From the independent MVAS provider point applications to be run. of view the operator is holding the strongest position as the operator has sovereign control over the customers by having direct billing relationship with them and controlling the access. BACKGROUND INFORMATION The term MVAS-provider used in a context of this paper refers to the mobile value-added service provider. With Content Content Application Application Portal / independence from the operator it is assumed that MVAS- Provider Aggregator Developer Provider Operator provider will offer services not only for the customer base of the specific operator, but for the broader audience and Figure 1: MVAS value change. often with its own brand. Even as the value chain has been criticised not revealing the real dynamics and structures of MVAS business SMS-Services [mob02] [buel02] [intv] this value chain presentation is Typical SMS services consist of service request sent as used in the several reports, and is used as a base for the an SMS to the short number of particular service pro- discussion here. vider. Typical examples of SMS based services are logo and ring tone purchases, chat and pushed news services (content is periodically pushed from the provider platform to the customer handset after service subscription). p Also more advanced services are provided in the area of support ro f mobile commerce, where SMS is used as a payment functions it method for services like car wash and parking payments, marketing

or even as a user-interface for specialised applications. aquisition oprations logistics delivery sales & content customer service WAP-Services Basic WAP-services offer browser like access to the mobile content. This content is textual and is sometimes supported with limited black and white graphics. WAP is also used as a user interface for the same services that are accessible via SMS in addition to more complex solutions Figure 2: Value chain applied to MVAS provider. like for example PIM, on-line banking and betting ser- Figure 2 presents a simplified view on the MVAS pro- vices. vider value chain based on the Porter’s [por91] model. Impacts to the support functions are left out of the scope of this paper. The five areas of the value functions are described in Table 1 from the MVAS-provider point of view.

75 Table 1: Value function descriptions.

Value Function Description Content acquisition Content acquisition is understood Operator A Network Operator B Network to contain not only the actual operator operator content (logos, news feeds, Java- gw gw games etc.) but also subcontracted applications for the content provisioning (Purchase Logistics, MVAS MVAS [por91]). gw gw Operations The service provisioning platform

and required technology MVAS Provider Service Platform development (Operations, [por91]). Delivery Logistic Connectivity required for service provisioning, including billing Figure 3: MVAS provider’s operator connections in system implementation traditional GSM network service provisioning. (Delivery Logistics, [por91]). Introduction of new network technologies will enable more Sales & Marketing Sales and Marketing (Sales and internet-like approach to the service implementation. If Marketing, [por91]). operators decide to work with “open-out” model instead Customer Service The Helpdesk functions and other of “walled garden” then access implementation from the customer support (Service, [por91]). MVAS provider point of view will change to follow the model presented in Figure 4 (solid line). The service-provisioning platform is a base for the business activity. How ever, the significant role is held by the Unfortunately, this approach often means restrictions in content, especially pricing and availability of it. At the availability of the customer identification information and middle of Figure 1 are service delivery and revenue col- billing functionality. In such model the service communi- lection. These two form a critical success factor for the cations is separated from the billing data exchange and independent MVAS provider because of two reasons: dedicated connections are again required (dashed lines in Figure 4). Other option is that MVAS-providers integrate First, MVAS operations are mass-business with possibil- into operator controlled service portals in order to get ity of customer churn: transaction peak-loads are signifi- functions like billing enabled. cant, and failure to provide the service is likely to drive customers away. Second, the small gross margin requires One option for the MVAS provider is to use other meth- revenue collection mechanism that is efficient, economic, ods for customer identification and presence information and suitable for the micro-payment transactions. Currently exchange. This is area where industry proposals for mo- operator capabilities play a major role on implementing bile personality and presence standardisation may prove both of these. to be interesting (Liberty Alliance Project [lap] and Wire- less Village [oma]). It needs to be pointed out, that especially in Liberty Alliance Project operators are likely to have strong position as Identity Provider. TECHNOLOGY

Network Connectivity Service interfaces for SMS are built per operator (see Fig- Operator A Network nas nas Operator B Network ure 3). Gateways are used for adapting to various SMS- protocols and per operator billing information exchange, Internet and the same approach is often applied also to the WAP connections. This reflects operator tendency to control the services available to its customer base (“open-in”, Service “open-out” and “walled-garden strategies”) [mait02]. Point Also, as exchange of customer identification and billing information takes place within the protocol an open net- MVAS Provider Service Platform work would introduce security compromises. In most cases MVAS provider buys capacity and connectivity together with the billing functionality from the op- Figure 4: Open service access via Internet. erator, and as a part of the deal operator takes care of the revenue collection.

76 Handsets Table 2: Connectivity changes.

The new breed of mobile handsets that is already mar- Technology Change Business Impact keted introduces colour displays with higher resolution and affordable price [nok] [soe] [sam]. This is anticipated Packet networks No implicit authentication. Need creating newly born interest to the WAP [intv]. And as to adapt new authentication and operators upgrade their networks browsing becomes more authorisation mechanism. flexible and economic way of accessing content and user Connectivity not controlled by interfaces with mobile handsets. Increased processing operator (if operators allow free power of handsets will allow specialised applications to network access). be run at the handsets. Faster rollout-span (if no operator specific solutions are required). RESULTS Operator billing may be based on data transfer. Distinction on operator price and Technology Impacts service price. The following major technology changes affecting to Costs harder to predict (customer MVAS environment are identified: point of view). Connectivity: Loss of billing connectivity with open network access is likely.· · From circuit switched networks to the packet networks Operator specific billing solutions · Higher bandwidth availability are likely. · Location information availability Shared access networks allow new kind of application logic (back Handsets: ground applications, connection is · Higher resolution and colour displays ”always on”). · Enhanced browser support (xHTML, other variants) Higher bandwidth Better user experience. · Terminal side add-on applications (Java2ME, Brew, New breed of applications (rich Symbian) content). · Enhanced personalising features: wallpapers, melodic Increased capacity needs for ser ring tones vice provisioning, network infra structure changes for MVAS · DRM provider. Both increasing network bandwidth and growing display Location Information New application opportunities. resolution can be seen as a continuum of development track. Introduction of different kind of browsers can also be seen as a continuum for the path that was started by Change of the network technology will leverage adapta- WAP. The change from circuit switched to the packet net- tion to more Internet-like business models where exter- works and the capability of adding terminal side software nal content and service providers come more important. to the handsets as a user maintainable operation can both Trend is even stronger due the economic slow-down of be classified as a discontinuous change in the technology the telecommunications sector and investment burden structure. These will impact to the services deployed and faced by the operators during the acquisition and ramp- the infrastructure alike. up of the 3G networks [sab02] [mait02] [buel02] [oll02] As the quality and the complexity of the content increases [intv]. also the costs for developing it will increase. In a same Even as operators will face new challenges, they still pos- way the end user prices of the services are likely to in- ses a very strong asset with existing customer relation- crease, and issues with copy protection and anti-piracy ships and related support functions like billing and mar- technologies will come more critical. Together with the keting. new content capabilities companies involved with mobile business have worked with the standardisation for the DRM [oma]. Handsets and Applications Connectivity and Network Table 3 below summarises most significant technological The Table 2 below summarizes most significant technol- changes on the handset side. ogy changes on the connectivity and network side with predicted impact to the MVAS business environment.

77 Table 3: Handset changes. ing). It’s quite uncommon that mobile services would be used for Internet surfing like activities [intv]. Technology Changes Business Impact It is expected that mobile services will require push from High resolution and More personalising features, the marketing side in order to gain popularity. Operators colour displays, more appealing and elaborate are aiming to secure their customer base (and revenue) in enhanced sound content. Higher content acquisi- the new market by building wireless service brands in capabilities tion costs. Growing interest to addition to the existing operator brand. Vodafone Live DRM. can be seen as an example of this [voda]. Features visible to the end user (customers). Value Chain Impacts Enhanced browser xHTML and other HTML The mobile business environment is changing from op- support variants allowing more browser- erator dominated simple value chain to more complex like user experience. value network. The change is driven by more complex Terminal side More versatility with the content service and content aggregates that will require more applications offering. specialised competencies. In such an environment it’s harder to keep dominant role and be able to respond to Dedicated client applications, the market needs. more internet like network traffic profiles. Operators still maintain few key assets that are hard to beat: existing micro-payment capable billing systems, IM and UM applications on existing billing relationship with the customer, and se- handsets. cure handset (and user) identification taking place on the New security considerations network level. This gives operators a strong position as (malicious applications). billing agent, trusted third party and payment provider. Digital Rights Standardised and in-build As the role of the operator will partially resemble those Management (DRM) DRM solutions. of traditional payment brokers in terms of financing the purchase and carrying the credit risk it is possible that such companies like Visa or American Express may also When content enhances the companies already providing see a business opportunity in the mobiles Internet. Sce- content in the Internet can utilise the existing offering in nario is interesting especially if operators allow open the wireless domain. Also the fact that revenue collection mobile Internet access, and MVAS-providers see opera- in the Internet has proven out to be difficult traditional tor solutions too complicated and expensive to implement. Internet service providers see mobile Internet as a media The scenario supports the assumption that operators will that offers built-in micro-payment environment. Also try to protect their assets as much as possible. mobile operators are aiming to guarantee the solid rev- As regulatory changes open up network access also enue streams [oll02] [intv]. This trend can be seen al- MVNO [mia02] are expected to appear in the 2.5G and ready as traditional Internet players are “mobile enabling” 3G network environments. Depending on the capabilities their communication services. Such companies as ICQ of these MVNOs they may offer an alternative to the ac- and MSN are all offering mobile enabled messaging ser- tual network operators as connectivity and billing pro- vices [icq] [msn]. vider. As the number of terminal side applications is rapidly in- With more demanding environments it’s expected that creasing and current state of the handset side application smallest mobile application providers will face difficul- security is not perfect (e.g. within Symbian platform a ties on financing the required operational cash flow. This malicious application have access to the dialling capa- is likely to will lead to new partnerships, mergers and bilities) a well-known and trusted brand comes an impor- acquisitions on the content creation side. Also the fact tant marketing asset. that companies currently offering Internet content and having interest on mobile domain will require new competencies is one factor that is likely to strengthen this trend. Internet and Mobile Convergence As the nature of the mobile service usage profile tends to As the combination of the network and handset is evolv- favour known brands and services the value of the brand ing the convergence of Internet and Mobile services is is likely to be an important asset. This suggests that the expected to take place into some extent, but differences role of the known content aggregators and service portals among use-profiles will have an impact to this phenom- will be important. enon. In the mobile world users are more interested (in addition to the phone personalization services) on the ser- From the independent MVAS provider point of view these vices that they are already familiar with (news services, changes have clear impact to the value chain. The im- existing SMS services), that offer instant gain (parking), pacted areas are summarised in Figure 5. The handset or into which mobility brings some value added (messag- changes together with connectivity changes will impact

78 practically into the whole revenue chain of the indepen- the MVAS-provider point of view highly depends on the dent MVAS provider: nature of the services provided. When limited to the simple content services targeted to the consumers, the role of customer service is likely to be negligible. With more complex service like corporate PIM the customer service may even work as a differentiating factor among the MVAS-providers. p support r o Figure 6 presents the expected change factors affecting fi functions t the traditional MVAS value chain as a whole. As can be seen, the whole value chain will be impacted. This as- marketing operations aquisition logistics delivery sales & content customer sumption is also supported by the literature [sab02] service [mait02] [buel02] [oll02] and the interviews [intv].

Content Application Content Application Connectivity Provider Developer Aggregator Provider Provider

handset connectivity changes changes more complex content new content providers handset applications portals internet existing operators platfrom enhancements payment companies MVNOs

Figure 6: Changes within the value chain. Figure 5: Impacts to the MVAS-provider value chain.

Content Acquisition: More elaborate content will require CONCLUSION enhanced skill sets from the content providers. This is The emerging technology changes in the area of the con- also likely to increase content acquisition costs and re- nectivity and handset technologies will enable clear change spective copyright holders become more demanding on in the mobile services and their provisioning, and the DRM. whole MVAS value chain will be affected. In the future, Operations: Together with the content and services also the concept of value chain will more likely resemble the the backend systems required for service provisioning value network where competence and value added are the need to change. The traffic profile starts to resemble key success factors. Internet service traffic profiles instead of traditional cel- As the technology will allow more and more internet like lular network SMS traffic profile. Handset diversity will content and user experience it’s expected that companies require handset recognition and automated content adap- that have been active on the Internet side will seek new tation in order to provide satisfactory user experience. opportunities on the mobile area, how ever, the direct Handset side client-application technologies will require adaptation of internet services will not work. automated version control. Operators who have held sovereign control over the ex- Delivery Logistics: Shared access networks are likely to isting 2G cellular domain will face a new situation where require new customer identification and authorisation they need to adapt to the new roles. Further, they need to methods. Until now customers have been reluctant to expand the mobile value chain with competence lead part- utilise traditional user-id/password authentication method nerships or adapt to the simple “bit pipe strategy”. for mobile services as they are used to rely on MSISDN providing explicit authentication and authorisation. It is The content will become more elaborate and require more likely that MVAS providers need to keep adapting into experience and investments to provide, favouring com- the connectivity provider (operators, MVNOs) specific panies already holding assets on the domain of on-line authentication and billing technologies. On the billing side content. And as the value of the content increases, DRM (revenue collection) it is possible that new players like comes more important. On content area the diversity of credit card companies and other transaction service com- the mobile environment is likely to push existing content panies (e.g. Paypal [pay] or other internet wallet solu- aggregators to expand their competence networks. tions) will enter to markets. Slipping into the Internet like business model where “free Sales & Marketing: In the mobile internet markets the content” would dominate the content offering is a threat. service discovery and visibility will be important. Also, However, this situation is unlikely, as the industry is show- as service prices are expected to rise the concept of a strong ing serious interest on keeping the micro-payment and brand becomes increasingly important. DRM model built-in to the mobile Internet model, and the existing Internet companies are reaching to the mo- Customer Service: The role of the customer service from bile side in order to enable billing mechanisms. Mobile

79 services may also concentrate to the services into which REFERENCES mobility brings so much value added that people are will- [buel02] Franz Buelling, Martin Woerter: Development Per- ing to pay for it. Further, it is possible to replicate appli- spectives, firm strategies and application in mobile commerce, cation models that people are already used to pay for (like Journal of Business Research, 5830, 2002. SMS messaging). [icq] ICQ, http://www.icq.com, March 2003. From the independent MVAS provider point of view the change in the business environment will impact the whole [intv] Interviews executed by author during 1Q2003. Opinions expressed in the interviews are professional opinions of the in- value chain and operating environment. The key success terviewed individuals and do not necessarily reflect the views factors are versatility to adapt to the heterogeneous envi- of the company. References to the interviews are always done ronment and technological changes with rapid pace, as to the summary of the trends, and no references to the opinions well as an ability to create competence networks with of an individual are made. content and solution providers. Capability if implement- Interviewees: ing efficient micro-payment channel is crucial and hav- Antti Jussi Suominen Vice President, ing good relationships with the operators is an advantage, M&A and Partnerships, although some other interesting micro-payment options Sonera Zed may appear to the markets. Jyri Hagman Director It is also interesting to ask what the future role of inde- End-to-end Business Programs pendent MVAS providers is. It may be that the current Nokia financial situation of the operators together with the rapid change of the technology has provided a window of an Paul Schulz Head of Mobile Technology Yahoo! Europe. opportunity that may close in the future as business environment matures. In such a world it may be that MVAS [lap] Liberty Alliance Project, http://www.projectliberty.org. services are not provided by an independent company but [mait02] Carleen F. Maitland, Johannes M. Bauer, Rudi as a part of other existing business like those that opera- Westerveld: The European market for mobile data: evolving tors and Internet service providers are running. value chains and industry structures, Telecommunications Policy 26, 2002, p. 485-504. In addition to that biggest single threat to the independent MVAS provider is that operators will adopt “closed gar- [mob02] Mobile Matrix AB: The saviour of 3G: the Leading den” strategies to the service offering. 100 Mobile Application and Content Providers – Strategies for Success, BWCS Ltd. 2002. [msn] MSN, Microsoft Network, http://www.msn.com, March 2003. ABBREVIATIONS [nok] Nokia, http://www.nokia.com, March 2003. [oll02] Phillip Olla, et.al.: A value chain model for mobile data ATM Automated Teller Machine service providers, Telecommunications Policy 26, 2002, pages 551-571. DRM Digital Rights Management [oma] Open Mobile Alliance, GPRS General Packet Radio System http://www.openmobilealliance.org, March 2003 IM Instant Messaging [pay] Paypal, http://www.paypal.com, March 2003. MSISDN Mobile Station International ISDN Number [por91] Michael E. Porter: Kilpailuetu, Gummerus 1991Finn- MVAS Mobile Value-added Service ish translation of: Michael E. Porter: Competitive Advantage – MVNO Mobile Virtual Network Operators Creating and Sustaining Superior Performance. PIM Personal Information Management [sab02] Hemant Kumar Sabat: The evolving mobile wireless value chain and market structure, Telecommunications Policy SMS Short Message Service 26, 2003, pages 505-535. UM Unified Messaging [sam] Samsung, http://www.samsung.com, March 2003. UMTS Universal Mobile Telecommunications System [soe] Sony Ericsson, http://www.sonyericsson.com, WA P Wireless Application Protocol March 2003. [voda] Vodafone, http://www.vodafone.com, March 2003.

80 MMS and WIM Technology and Business

QINGHAI LI DEPARTMENT OF COMPUTER SCIENCE HELSINKI UNIVERSITY OF TECHNOLOGY [email protected]

ABSTRACT work in the UK [15]. Each short message has at most 160 characters. SMS has three major elements: mobile termi- The general descriptions about Wireless Messaging Ser- nal, wireless network and Short Message Service Center vices including SMS, EMS, MMS and WIM (Wireless (SMSC). SMS is a ‘store and forward’ mechanism, which Instant Messaging) are presented. After that, the overview means if the receiver is not available at the moment, the of MMS and WIM infrastructure is described and ana- SMS will be stored into SMSC and then deliver it once lyzed. The value chain and pricing strategy are important possible. so that we examine the roles and players sitting in MMS value chain and anticipate the income percentage of them. Due to the limitation of SMS, Enhanced Messaging Ser- There are two basic pricing models for MMS, as described vice (EMS) comes out as the complement to SMS. EMS sub-chapter 5.2. Finally, the MMS market predication, provides the ability to deliver the combination of simple including MMS traffic volume, MMS handset and MMS melodies, pictures, sounds, animations, and formatted text. subscriber, is illustrated by some figures. However, EMS is not running as popular as expected. One reason is the main mobile manufacturer, Nokia, ig- nores EMS and go directly to MMS. Another reason is 1 INTRODUCTION that MMS comes into reality. MMS has evolved from SMS (Short Message Service) MMS is able to send and receive the combination of text, and EMS (Enhanced Messaging Service. SMS gained sounds, images, and video to MMS capable handsets. great success though it provides the limited wireless mes- MMS features can be fully realized on the coming 3G saging services. EMS comes after SMS for improving the (3rd generation) wireless network. However, you needn’t capability of wireless messaging service. EMS can pro- wait to use MMS because of 3G network, since most of vide relatively richer messaging service like simple pic- MMS features are enabled in 2.5G (also called GPRS) ture and audio, but it isn’t an ultimate solution. Due to the and EDGE which are available nowadays. introduction of GPRS, EDGE, 3G networks and more capable mobile handset, it’s possible to re-construct wire- 2.2 Introduction to WIM less messaging technology from its infrastructure and provide more features to wireless messaging service. Under There are many desktop Instant Messenger (IM), How- this context, Multimedia Messaging Services (MMS) ever, those Instant Messengers normally can’t communi- comes out to suit the needs. cate with each other due to the lack of standard. Wireless Instant Messaging (WIM) is a popular topic Wireless Instant Messaging (WIM) has the similar nowadays. The desktop IM is common for instant chat- functionalities as desktop-based Instant Messaging ser- ting, but suffered with the lack of standards and unable to vices. WIM system is the client and server model. The roam. WIM solves these problems instead. client is connected with WIM server and two clients can talk with each other via server. Each client side can maintain the buddy list, know the 2 CONTEXT OF MOBILE MESSAGING presence status, chat instantly with each other, and search friends etc. In addition, WIM will also enable the rich messaging like pictures and audio. File sharing and Group 2.1 General description features are promised by Wireless Village IMPS specifi- Mobile messaging is closely related with this evolution cation too. process of wireless network. In the late 80s and early 90s, Currently there are two major standards which are under the first generation wireless network emerges. At this stage, developing currently: there is no mobile messaging. - Wireless Village initiative (http://www.wireless- In second generation (2G) of wireless network, the first village.org), a community established by Nokia, Ericsson, short message is sent in December 1992 from a Personal Motorola in April, 2001. Computer to a mobile phone on the Vodafone GSM net-

81 2.3 Introduction to MMS

2.3.1 MMS Content The Multimedia Messaging Service (MMS) is as its name suggests the ability to send and received messages com- prising a combination of text, sounds, images and video to MMS capable handsets. [14] MMS can be understood as “multimedia PowerPoint slide presentation” based on a common timeline [12]. The content types that MMS supports: Text. Unlimited text which can be formatted. Text with images, video, audio etc. Graphic. Graphs, tables, charts, animated GIF etc. Audio. Music, speech, streaming sound. Figure 1: Four main desktop IM Video. A segment of video clips, or streaming video.

Figure 3: Example of Using MMS

2.3.2 How MMS Works? As Figure 4 shown, two MMS capable terminal can not Figure 2: Wireless Instant Messenger example (Source: only send and receive MM (Multimedia Message) with wireless-village.org) each other, they can also get MM from Internet, for example, from email. If the mobile device doesn’t support MMS, the user will get an SMS message containing an URL where the message is stored. - Jabber (http://www.jabber.org), an open XML protocol standard for the real-time exchange of messages and presence information. Both standards are basically ready and they offer the same interoperability. Wireless Village is more competing com- Internet pared with Jabber because Wireless Village is founded Multimedia MMS by Nokia, Ericsson and Motorola which are the leader of Messaging Terminal Center wireless worlds. Furthermore, Wireless Village provides A (MMSC) the more features than Jabber like group chatting and file sharing.

There are two kinds of WIM clients. One is the built-in software in mobile terminal. Another is downloadable Old MMS Mobile software for mobile terminal. For downloadable WIM Terminal clients, mostly mobile terminal need support java and B Handset Jabber under J2ME platform. The downloadable WIM clients have worse usability than the built-in ones since they can’t access the mobile handset profile information Figure 4: The way of MMS communication (i.e. on call, available etc), and the downloadable WIM MMS is not delivered in real time as SMS. Both MMS client has to be launched separately and can’t be running and SMS follow the “store and forward” method. For under background.

82 MMS delivery, firstly, the sender sends a message to the 3 MMS MESSAGING ARCHITECTURE MMSC. When MMSC gets the message, it sends confir- MMS terminal – MMS client side device used to create, mations of “message sent” to the sender. MMSC then manage, send, receive and render the Multimedia Mes- notifies the recipient that a message is waiting. If the re- saging Service. cipient is not available at the moment, the message will be stored in MMSC. Otherwise, the recipient can down- MMS proxy relay – System element mainly responsible load the message immediately or download it later. As for interacting with MMS terminal. MMS proxy relay also long as the receiver downloads the message, the sender stores the MMS to the MMS server and communicate with will get a “message delivered” notification. other MMS proxy relays. MMS Server – The element providing the storage ser- 2.4 Comparison of MMS and SMS vice for MM messages. It can be combined with MMS proxy relay.

2.4.1 Content, Size and Features Email Server – Standard Internet email services. MMS is not theoretically limited by its size. Compared Legacy Wireless Messaging Systems – The current and with SMS which is usually limited to 160 characters, MMS previous Wireless Messaging services like SMS and pag- is much more size flexible. SMS content is the pure text ing. message; MMS could not only contain text, also audio, MMSM – Interface between MMS Client and MMS Proxy video, picture, animated cartoon etc. Even real-time video Relay. See section 3.3. stream may be supported. MMSR – Interface between MMS Proxy Relays of dif- MMS can be personalized according to the user’s profile ferent MMS systems. No standard specification defined stored in WAP gateway. Based on the user’s profile, MMS yet. See section 3.4. system can decide when to send Multimedia messages and how to present them. This feature is not supported by MMSs – Interface between MMS Server and MMS Proxy SMS. Relay. No standard defined yet.

2.4.2 Environment 3.1 MMS Terminal to MMS Terminal Messaging SMS appears in 2G environment and is still available in 2.5G (GPRS) and will exist in 3G networks (Mainly UMTS). MMS is not binded with 3G. MMS emerges in two phases. First phase is based on GPRS (2.5G) [12], rather than 3G. Once 3G comes to reality, more complicated features will be introduced. The storage is never a problem for SMS because the size of SMS is fairly small and SMS is stored only when the receiver is not available at that time. MMS is larger and it is stored in MMSC for downloading. In addition, the user may like to store the MM somewhere for long time, which is a business opportunity for the storage companies. MMSC is composed of many elements, which can come from different vendors. SMSC is monolithic.

Figure 6: Terminal to Terminal within one home system [17]

Figure 6 is the simplest case of Multimedia Messaging Service that two subscribers are in the same home network. MMS terminal A sends one MM to terminal B. First, the user activates his MMS client program. Then, he/she enters the target address. After that, he/she composes the Figure 5: Table comparison of SMS & MMS (Source: Multimedia message. Once the user asks the MMS client Mobile Stream) [14] program to send this MM out, MMS client will handle the MM and submit it to its related MMS Proxy Relay

through MMSM interface. MMS Proxy Relay receives the Multimedia Message, resolves the target address from

83 MM and finds it is in the same network. MMS Proxy Relay contacts with MMS server via MMSS interface and stores the MM in MMS server, then sends the notification to the MMS terminal B. MMS terminal B retrieves MM from MMS Server and notifies the sender that the MM is got- ten. Finally, Terminal B displays the received MM to the user.

Figure 8: MMS Terminal, Email Server and Legacy Wireless System [17] Figure 7: Terminal to Terminal between different home The connection between MMS Proxy Relay and Email system [17] Server should work in both directions: As Figure 7 shown, Terminal A and B are sitting in differ- - Sending MM to Email Server via Internet. By using ent network. A sends MM to B. As Figure 6, MMS Proxy SMTP (Simple Mail Transfer Protocol), the MMS Proxy Replay gets MM from terminal A, resolves the target ad- Relay submits the message to the destination host. In or- dress, but it won’t save MM into MMS Server in network der to pass by the specific MM format message through 1, it routes the MM to the target MMS proxy Replay in Internet, the message traffic needs to be converted into network 2 through MMSR interface. MMS Proxy Relay standard MIME format (Multipurpose Internet Mail Ex- in network 2 then contacts with its MMS Server (in net- tensions); hence the various media like video, audio and work 2 in this case), which saves MM into its storage. picture could be smoothly flowing through Internet. In After that, MMS Proxy Relay sends the notification to particular, the MMS special headers will be converted MMS terminal B. The left process is just same as men- into the proper headers with prefix of “X-Mms-“. The tioned in Figure 6. prefix of “X-Mms-“ will identify the message is “MMS” instead of others to the recipient, so the recipient can prop- 3.2 Communication among MMS Terminal, Email erly render this Multimedia Message. Server, Legacy Wireless System - Receiving messages from Internet Email System. The In Multimedia Messaging System, it is important to have MIME message from email system will be converted into compatibility with the early Wireless Messaging systems MMS format. Any headers with “X-Mms-“ are also con- like SMS plain text messaging. MMS device is able to verted back into MMS headers. handle the legacy wireless messaging, but not in reverse - Fetching messages from Internet Email System. MMS way. Due to the limitation of a certain GSM handsets, client can retrieve messages from Internet Email Server MM from the sender can’t be correctly processed. Under via MMS Proxy Relay by using POP or IMAP protocols. this situation, for example, an URL pointing to the MM will be sent to the receiver.

84 3.3 Interfaces between MMS Client and MMS tem. Proxy Replay · CSP – Client Server Protocol. It provides the access for Embedded Client within mobile terminal and desktop clients to communicate with Wireless Village Server. · CLP – Command Line Protocol. It provides the access for an Old Mobile Terminal to communicate with Wireless Village Server. · SSP – The Server-to-Server Protocol. It provides the access between different servers in same domain or different domains. · SMCNP – Server Mobile Core Network Proto- col. It gives the access to the Mobile Core Network so that Wireless Village server can get presence information and service capability information from network. Figure 9: Logical Architecture of MMS Client to MMS Proxy Relay (source: WAP MMS Architecture 2001, · CLI Client – Command Line Interface Client WAP Forum) using text message. · WV Embedded Client – Wireless Village Em- The relationship between the MMS client and MMS bedded Client. It is an embedded client inside a mobile Proxy-Relay is shown as figure 9. MMS Client is acting terminal. Different vendors could have different embed- as the MM sender/receiver. MMS Proxy-Relay is acting ded client with different functions and appearance. Gen- as an Origin Server(Pull operation) or a Push erally, WC Embedded Client not only provides the text Initiator(Push operation). The traffic between MMS Cli- message, also picture, audio and video are possible. ent and WAP Gateway is transferred using Wireless Trans- port Protocol – WSP (Wireless Service Protocol). Using of WAP WSP for transport doesn’t mean the use of WAP 4.1 Big Picture of IMPS System browser; instead, SMIL (Synchronized Multimedia Inte- gration Language) is used as the method of MMS presen- Wireless Village, founded by Nokia, Ericsson and tation layer. WAP Gateway provides the standard WAP Motorola, produced “IMPS, Instant Messaging and services including HTTP methods and PUSH services. Presence Services” specifying a set of universal specifi- WAP Gateway encapsulates the messages as HTTP and cations for mobile instant messaging and presence ser- transfers them to MMS Proxy-Relay. vices among mobile devices, mobile services and Internet- based instant messaging services. Similarly, the traffic from MMS Proxy-Relay to MMS client is in reverse way.

3.4 MMS Presentation The concept of MMS presentation means the ordering, layout, sequencing and timing of multimedia objects on the terminal screen and other devices such as a speaker. There exist two important languages for MMS presentation - WML (Wireless Markup Language) and SMIL (Syn- chronized Multimedia Integration Language, XML based language). WML support the layout and sequencing capabilities as with browsing. Compared with WML, SMIL provides the extended functions for presentation like timing and animation. SMIL is similar with HTML in syntax and structure. It is an essential way of presenting rich, interactive multimedia content based on common timeline.

4 WIM SYSTEM ARCHITECTURE Wireless Village and Jabber provide the similar wireless instant messaging standards. In this chapter, we will concentrate on the architecture of Wireless Village WIM sys- Figure 10: IMPS System Architecture [18]

85 IMPS is the client-server structure. Wireless Village Server service features it can use. is the center point of IMPS system. User Profile Management describes how the user wishes As Figure 10 shown, if a client (either CLI Client or Em- to manage and interact with their communication services bedded Client) is going to talk with Wireless Village Server [18]. via CSP or CLP protocols, it can’t access the Service El- Service Relay is used when server-server communication ements directly. Service Access Point handles the request happens. It routes all of requests and response among serv- from the client side. Service Access Point firstly authenti- ers through the Server-Server Protocol (SSP). cates the client, then, it uses Service Discovery mechanism to find out what services are available for the client. After that, Service Access Point will read the User Pro- file to decide, for example, what services are allowed to 5 MMS VALUE CHAIN AND PRICING be used by the client, terminal capability (i.e., support audio or video), client interface and so on. The most im- 5.1 MMS Value Chain and Players portant aspect is Presence Service. Service Access Point will read the presence information in user profile if the There are five players in MMS value chain. user sets it explicitly. If not found, then Service Access Point will access the Mobile Core Network to get the terminal presence status (powered off, on call, available…). Figure 11: MMS Value Chain (Source: Siemens Mobile If the user is going to talk with another user, Service Ac- Acceleration 2002) cess Point will start Instant Messaging Service. - MMS Infrastructure Provider: providing basic messag- From Figure 10, if two users are located in different doing infrastructure. For example, MMSC provider. mains, Wireless Village Servers in different domains need to communicate with SSP (Server to Server Protocol). - Middleware Provider: the provider for application and content adoption platforms & gateways. I.e. the companies providing gateway between mobile network and 4.2 Service Elements Internet. There are four Service Elements that are accessible - Application Provider: end user messaging applications, through Service Access Point - Presence Service Element, like MMS composer, editor, etc. Instant Messaging Service Element, Group Service Ele- ment and Content Service Element. - Content & Service Provider: providing content and messaging value-added services like text, picture, audio and The Presence Service Element provides functionality for video etc. presence information - Operator: providing mobile network, MMS transport, Management [18]. The functions includes modify, set and and billing services. get presence information. Instant Messaging Service Element provides the features of sending and receiving instant messages. Infrastructure Middleware Application Content & Operator Service Provider Group Service Element is used to manage the user group. ______15 % About 5% About 1% About 0 - 35% 79 - 44% Content Service Element provides the feature to share the content like files with other users. Table 12: value percentage of roles (source: Siemens th 4.3 Service Access Point Mobile Acceleration, Nov, 6 2002) Service Access Point provides the interacting interface As Table 12 shows, the biggest benefit is from Operator between WV Server and its environment. It contains the side. Content & service provider shares 0 – 35% benefit interface to the clients, interface to other WV servers, in- value, depending on the attractiveness of content and how terface to Mobile Core Network, and Interface to other MMS content is charged. non-WV servers. Service Access Point mainly provides four functionalities: 5.2 Pricing Authentication and Authorization, Service Discovery and Service Agreement, User Profile Management and Ser- The reasonable pricing level is not clear yet, but a generic vice Relay. model suggests that a MMS must not cost more than a regular postcard, about Euro 1 (postcard incl. stamp) [4]. Authentication is used to verify the identity of an entity, One strategy is lower price for entry and tries to increase i.e. user, network or application. Authorization is aimed the price for high value content. at checking what an authenticated entity is allowed to do. There are two most probable pricing models, as follow- Service Discovery allows the applications to identify what ing:

86 Figure 13: pricing models (source: following ICM NMA)

Complicated pricing only makes the customer confused and blocks the customer’s motivation for using MMS, so keep it simple. The following table shows the practical pricing in real Figure 15: Percentage of Mobile Traffic Volumes (Source: life from different companies of different countries. For ARC Group Mobile Services & Applications Industry Price per Message model, the price range is from 0.26 Survey 2002) Euro to 1.35 Euro. The most expensive price for each MM is 1.35 Euro from Telenor of Norway. The cheapest In order to make MMS successful, there should be enough price for each MM is 0.26 Euro from PTC Company of subscribers registered to MMS services. According to the Poland. big success of SMS which is much less features less than MMS, MMS is obviously more attractive to the user. From Frost & Sullivan investigation, MMS subscriber will be continuously growing, as Figure 16 shown.

Figure 14: Pricing example (Source: Siemens Mobile Acceleration, Nov, 6th 2002)

Price per Message has the limitation of traffic volume. The price is different within different range of traffic. Same as Flat Fee per Month, the traffic per month is restricted Figure 16: MMS Subscribers in Europe, 2001 – 2006 to a certain level. [2]

The success of MMS is not independent from other fac- 6 MMS MARKET EXPECTATION tors. It heavily depends on the roles of its value chain. Any unexpected problems in its value chain will cause Operators, especially in Europe and Asia, are starting to the failure of MMS. deploy commercial MMS with a view to secure growth. As these initial services are launched, companies from all First, Mobile handset is one key of MMS system. Cur- parts of the mobile value chain are betting that MMS will rently MMS capable phone doesn’t reach large number revitalize their service offerings and enable the much-an- of users. MMS features can be fully functioning in MMS ticipated next-generation of growth. MMS is the first ser- capable phones instead of legacy ones, so the popularity vice for 2.5 and 3G data networks to be aimed at the mass of MMS depends on MMS capable phones. market, Mobile operators especially are hoping that their considerable investments in packet-based network infrastructure and third-generation licenses will be justified by MMS.

87 8 REFERENCES [1] Arcgroup.com, Multimedia Messaging Services, December 2002, http://www.arcgroup.com/homepage.nsf/tag/mms [2] Eduardo Gonzalez, European Multimedia Messaging Ser- vices Market (MMS), 2001, http://www.frost.com/prod/ brochures.nsf/474fdcdb0941e35080256b770074c3e4/4e03aa 6229a4c10c86256c5d0059cfef/$FILE/MMS%20Virtual %20Brochure.pps [3] iBilt Technologies Ltd., GPRS A NEW HORIZON, 2002, http://www.ibilttechnologies.com/mms.pdf [4] Ingo Potthof, Multimedia Messaging Service(MMS) – New th Figure 17: MMS Handset Penetration (Source: Business Opportunities for the Mobile Industry, November 6 2002, http://www.telecom2002.co.il/presentations/Ingo% Mobilestreams) 20Potthof.ppt According to Mobilestreams, the market of MMS hand- [5] Kimmo Kullervo Pekkola, Market Introduction of A Mobile set will smoothly grow up, but the mass market won’t be Data Service: Description of The Value Network, May 24, 2002. reached before 2004. [6] Kobi Marenko, Partnership as a key success factor in the In addition, the quality of service and content, pricing, MMS market, 2002, http://www.telecom2002.co.il/presenta- and billing all affects the popularity of MMS service. tions/Kobi%20Marenko.ppt In spite of its challenge, no one doubts the large profit [7] LogicaCMG, LogicaCMG powers KPN Mobile’s provision of multimedia messaging services (MMS), March 24, 2003, potential of MMS. More and more operators started or http://www.logica.com/company_info/news/press_releases/ are starting MMS service, as shown in Figure 18: press_releases.asp?display=detail&id=658 [8] Logica, The essential guide to Multimedia Messaging, http://www.logica.com/pdf/telecom/Mmsguide.pdf [9] MessageVine, MessageVine IM Server, http://www.messagevine.com/products/instant_messaging.htm [10] Mobileinfo, Technology Logica and Mapflow Demo First Wireless MMS Location Services, Feburary 2002, http:// www.mobileinfo.com/News_2002/Issue08/Logica_Map flow.htm [11] MobileMMS.com, MMS Devices, http://www.mobilemms.com/devices.asp [12] MobileMMS.com, MMS FAQ, http://www.mobilemms.com/mmsfaq.asp [13] MobileMMS.com, MMS Applications : Content Sample, Figure 18: MMS operators (Source: ICM NMA and Glo- April 2003, http://www.mobilemms.com/sample_mmsapps.asp bal Mobile/GSMA) [14] Mobile Streams Ltd, NEXT MESSAGING An introduction to SMS, EMS and MMS. September 13, 2001. http://www.mobilewhitepapers.com/pdf/messaging.pdf 7 CONCLUSION [15] Mobile Streams Ltd, Success 4 SMS White Paper, Febru- Multimedia Media Messaging and Wireless Instant Mes- ary 2001, http://www.mobilewhitepapers.com/pdf/SMS.PDF saging are just in startup phase. Currently MMS and WIM [16] Roman Vichr, Tips&tricks: Multimedia Messaging Ser- content are not rich enough; mostly text and simple audio vice, September 2002, http://www-106.ibm.com/developer and video are supported. Therefore, we need more MMS works/library/wi-tip25.html?dwzone=wireless and WIM content and applications providers. The MMS [17] WAP Forum. Wireless Application Protocol Multimedia infrastructure is not complete ready and popular in op- Messaging Service Architecture Overview Specification. April erator sides. Some standards and interfaces need to be 25, 2001, http://www1.wapforum.org/tech/documents/WAP- settled down. Wireless Instant Messaging standards were 205-MMSArchOverview-20010425-a.pdf established mainly by Wireless Village and Jabber, but it is not clear who will win finally. For mobile terminals, [18] Wireless Village, “Wireless Village The Mobile IMPS Ini- tiative, System Architecture Model Version 1.1”, 2001, http:// there exist a few MMS and WIM capable devices, but not www.openmobilealliance.org/wirelessvillage/docs/WV_ enough yet. More MMS and WIM capable functions need Architecture_v1.1.pdf to be added in. For network, 3G is not ready yet, so bandwidth and some features of WIM and MMS can’t be realized though GPRS and EDGE are available to support most functions of MMS and WIM.

88 Mobile Entertainment

MIN CHEN COMPUTER SCIENCE DEPARTMENT HELSINKI UNIVERSITY OF TECHNOLOGY [email protected]

1 INTRODUCTION meeting the ever-growing needs of their customers worldwide. Mobile entertainment is a rapid growing segment of the mobile communications industry and evolves into the leading value-added mobile services. Mobile entertainment 2.2 Short Message Services (SMS) refers to entertainment products run on wireless net- The SMS is a service that enable mobile subscriber to worked, portable, personal devices. The term “mobile send and receive short text messages on their mobile entertainment” excludes mobile communications like per- phones. The messages can be received at any time even son-to-person SMS, voicemail, and mobile commerce during data or voice transmission. The first short mes- applications. sage was sent in December 1992 from a PC to a mobile The mobile entertainment services utilize Internet tech- phone in the Vodafone GSM network in UK. Over the nologies, mobile phone technologies and popular enter- past few years, the volume of SMS messages has increased tainment contents to offer amusement to subscribers. significantly. This implies the immense value of service Mobile games, ringing tones, pictures, and music are ex- provision and revenue generation. amples of the type of mobile entertainment services already existing today. 2.2.1 SMS Applications The main purpose of this report is to analyze the usability of supporting mobile entertainment technologies and rel- The vast majority of SMS usage is accounted for by con- evant market issues. The information in this report is fo- sumer applications. These applications are becoming ever cused specifically in the field of mobile entertainment more abundant, including simple person-to-person mes- services. It is beyond of the scope of this report to cover saging, voice mail, fax, unified messaging, games, every aspect of the supporting technologies and related ringtones, pictures, voting, gambling, information ser- business issues. vices, mobile banking etc.

Games 2 TECHNICAL ANALYSIS Games are a potentially huge service involving interactivity that many people see as being a key applica- 2.1 GSM System Overview tion for future mobile networks. Taking the i-mode service from NTT DoCoMo as an example, over half of the total usage volume is orientated to entertainment services The GSM (Global System for Mobile Communications) with a quarter of i-mode users subscribing to games ser- system is currently a globally dominant system and as the vices. Other services that are popular include ringtones, basis for approximately 70% of future 3G networks. GSM pictures, and horoscopes. [4] provides a common set of compatible services and capabilities to all mobile users across Europe, as well as countries in Africa, Asian, Australia, and Latin America. The Pictures Third Generation Partnership Project (3GPP) now gov- Picture messages are quickly gaining popularity, repre- erns the standard. GSM is a growing and evolving system senting the earliest stage of mobile evolution to a more that offers expanded services. GSM networks will con- visual interaction. The images are simple bitmaps. tinue to dominate and eventually evolve into more sophisticated 3G networks. The GSM family consists of today’s GSM, General Packet Radio Service (GPRS), Unified messaging Enhanced Data rates for GSM Evolution (EDGE), and The ideal of Unified Messaging is to provide a single in- third generation GSM services (3GSM). Together the terface for accessing various messages such as fax, voice GSM family provides a flexible platform and an evolu- mail, short messages, email etc. tionary path that enables network operators to offer a potentially limitless range of revenue generating services,

89 Ringtones weather reports, financial market information, and sports scores are available. The introduction of push has im- Custom ringtones downloaded via SMS are becoming proved the environment for m-commerce services. Glo- increasingly popular. Ringtones are the tunes that the bal Positioning System (GPS) functionality with 2.5G will phone plays when someone calls it. The service enables facilitate location-based services. phone users to be able to change their ringtone to distin- guish from others. The most popular ringtones varies very Entertainment services are expected to be a major appli- considerably but tend to be popular music, television or cation of WAP technology. Current offerings are limited film theme tunes. and distributed selectively. Enhanced services are likely to include multimedia, MP3 and other data downloads. 2.2.2 Usability Analysis 2.3.2 Usability Analysis SMS user interfaces are tedious and difficult to manipu- late. The MMIs (Man Machine Interfaces) are not stan- The user interfaces for WAP devices remain tedious and dardized, resulting in varying usability among handset difficult. The browsing capabilities are limited. Although interfaces. Content delivered via the SMS medium are WML and WMLScript code is standardized, it is unavoid- generally text based, but can use also graphics where ser- able that different micro-browsers interpret applications vice providers have installed the 8-bit Data Coding differently. Testing WAP applications on various hand- Scheme. In each message, content is restricted to 160 sets can reduce the problem, but cannot eliminate it un- character of Latin alphabets, or 70 characters of non-Latin less applications are optimized for all handsets and the alphabets. Text layout cannot be controlled effectively due server configured to recognize each one. Connection fail- to differences in screen sizes among various mobile de- ures sometimes occur in spite of apparent full network vices. Minimizing transactions with the content server, coverage, which restrict access to services. Moreover, the configuring servers for maximum efficiency is necessary. delivery of effective, enjoyable entertainment services depend much on WAP bearers. Combining various hand- SMS is a commercially important technology at present sets with various gateways currently in use can yield un- for a variety of applications. A sound business case exists predictable performance. However, compare to SMS, for SMS games at present despite inherent restrictions richer content is possible with WAP because WAP micro- due to the technology’s immense popularity and user base. browser can support 15 characters for titles and 750 char- However, a user-centered design approach is essential to acters for page text. Presentation remains crude, but al- ensuring the commercial viability of SMS entertainment lows for better visual quality overall. And transactions on services. As the presence of other more flexible technolo- WAP are faster than with SMS. gies improve the richness, reliability, connection speed, and possibilities for mobile entertainment, the relevance WAP 2.0 will facilitate improvements at service develop- of SMS will decline. ment and deployment. The J2ME handsets will facilitate extended functionality and personalization. However, to achieve optimal performance, J2ME applications require 2.3 Wireless Application Protocol (WAP) more sophisticated handsets and faster 3G networks sup- The Wireless Application Protocol is a de facto world port. Optimized 2.5G networks and handsets represent standard for the presentation and delivery of wireless in- better opportunity for mobile entertainment. However, 3G formation and telephony services on mobile phones and networks and compliant handsets represent the most sub- other wireless terminals. [7] The purpose of the WAP is stantial opportunity. to empower mobile user to access and interact with information and services easily and fast. 2.4 3G Technologies In June 1997, Ericsson, Motorola, Nokia, and Phone.com Third-generation technologies will supplant current tech- founded the WAP Forum. The WAP Forum has drafted a nologies and 3G systems are expected to offer better sys- global wireless protocol specification for all wireless net- tem capability and higher data transmission speed to sup- works, various industry groups and standards bodies. This port wireless Internet access and wireless multimedia ser- specification enables manufacturers, network operators, vices. The services include high-quality voice, Internet, content providers and application developers to offer com- and multimedia services. patible products and secure services on all devices and networks, resulting in greater economies of scale and universal access to information. [7] The WAP specification 2.4.1 3G Network Standards extends and leverages existing technologies, such as IP, HTTP, XML URLs etc. International Mobile Telecommunications 2000 (IMT- 2000) is the ITU’s proposed standard for 3G networks. There are two major standards, Wideband Code Division 2.3.1 WAP Applications Multiple Access (W-CDMA) and Code Division Multiple Access 2000 (CDMA2000). Current WAP applications include unified messaging and e-mail services. Information services delivering news,

90 The WCDMA is the European Telecommunications Stan- 2.4.3 Applications dards Institute’s choice for 3G cellular systems. One of 3G applications will substantially increase wireless func- the main advantages of WCDMA is that it can extend the tionality and richness of content over previous genera- existing GSM core network, maximizing reuse of exist- tions. These applications include communications, infor- ing equipment. It can also overlay TDMA and CDMA mation, entertainment, and m-commerce. networks. The implementation of WCDMA will be a technical challenge due to its complexity and versatility. Entertainment applications are likely to achieve unprec- WCDMA supports theoretical data rates up to 2Mbps, edented usage. Gaming is a key application with regard though only supports 384kbps for wide area access. The to entertainment and location-based services will create a much-enhanced data rates require a 5 MHz wide carrier. realistic gaming environment. Audio and video applica- The capacity of WCDMA is 50-80 voice channels per tions are also expected to be popular. Multimedia Mes- carrier. The Implementation requires also compliant hand- saging Service (MMS) facilitates richer and more per- sets. This standard is expected to account for approxi- sonalized content. Unified messaging will provide a single mately 70% of 3G networks. interface for downloading communications from various sources. The CDMA2000 specification was developed by the Third Generation Partnership Project 2 (3GPP2), which consists of five telecommunications standards bodies includ- 2.4.4 Usability Analysis ing ARIB and TCC in Japan, CWTS in China, TTA in Korea and TIA in North America. CDMA2000 has al- Compared with the 2G systems, 3G systems offer better ready been implemented to several networks as an evolu- system capacity, high-speed wireless Internet access, and tionary step from cdmaOne as it provides backward com- wireless multimedia services including audio, video, im- patibility with the IS-95 architecture. CDMA2000 can ages and data. The user will experience vastly improved support mobile data communications at speeds ranging interactivity and visual impact. 3G devices will be im- from 144kbps to 2Mbps. CDMA2000 1x EV-DO and proved with enhanced battery capabilities, larger screens CDMA2000 3x are ITU approved 3G standards. The 1x and higher resolutions, more sophisticated mini-brows- Radio Transmission Technology (RTT) uses 1.25MHz of ers, more memory, and greater storage capacity. Enter- bandwidth to support data rates of 144kbps and up to tainment applications will become streamlined and viable 307kbps. 1xEV-DO supports max 383kbps to 2.4Mbps. across various wireless devices. To ensure application 3xRTT will use 5MHz of bandwidth, support data rates integrity and client safety, security issues need to be ad- up to 2Mbps. This standard is expected to account for dressed. However, 3G entertainment services becoming approximately 13% of 3G networks. popular require sufficient penetration of 3G networks and compliant devices.

2.4.2 Enabling Technologies Fundamental improvements in network implementations 3 MOBILE ENTERTAINMENT MARKET together with sophisticated supporting technologies such ANALYSIS as Mobile Station Application Execution Environment (MExE) and Bluetooth will enable richer multimedia ap- 3.1 Mobile Entertainment Business of Today plications. There are regional differences in the case of mobile en- MExE is a specification managed by 3GPP that will stan- tertainment. The Far east, specifically Japan and Korea, dardize the environment for executing mobile applica- is considered to be the leading region for uptake and actions across the full range of 3G devices. It incorporates ceptance of mobile entertainment. Europe is second in WAP and Java in a framework specifying a significantly acceptance with Scandinavia, the UK and Italy being the more flexible and secure application for 3G handsets. [2] most aggressive markets. The US, as far as the overall Bluetooth is a de facto standard enabling wireless con- adoption of mobile services goes, trails when it comes to nectivity between devices such as mobile phones, mobile adoption of mobile entertainment services. Market ana- computers, and PDAs. It was conceived by Ericsson in lyst reports are still forecasting multi-billion dollar indus- 1994, and is now developed by the Bluetooth Special In- tries for mobile games, mobile music and other forms of terest Group. Bluetooth provides low-cost, low power, mobile entertainment. [5] robust, secure, efficient, high capacity, ad hoc voice and The most successful service to date is ring tones. Close data networking. Bluetooth requires a low-cost transceiver behind the success of ring tones are mobile games. The chip be installed in each device. The transceiver utilizes industry is expected to grow from 1.5 billion Euros in the unlicensed frequency band of 2.4 GHz that is avail- 2001 to 15.4 billion Euros in 2005. [1] Although the vast able globally to transmit and receive data. The connec- revenues are generated by mobile entertainment, insuffi- tions between devices can be point-to-point or multipoint cient amounts of these revenues flow through to the com- in maximum range of 10 meters. Data can be exchanged panies providing the content and technology. at a rate of up to 1 Mbps.

91 3.2 Revenue Models Motorola, Ericsson and Siemens etc.

Consumers (Co) 3.2.1 Mobile Entertainment Value Chain Consumers use the services provided by above service As illustrated in Figure1, the service chain for mobile providers and are also the resources of the revenue for entertainment industry typically looks like this: services providers.

There are also many variations of this chain. Many com-

CP Op D Co panies in this chain may occupy more than one role. AD SP PP Pub M

3.2.2 Revenue Sharing A logical and typical model for mobile entertainment com- Figure 1 Mobile Entertainment Value Chain panies is based on revenue sharing with mobile operators. Operators can easily take the advantage of existing Content Provider (CP) customer relationships to market entertainment services to both existing and potential subscribers. A user plays a Content providers create original contents or provide SMS game is usually engaged in at least 10 transactions, popular brands, characters or themes for mobile enter- which can generate significant revenue for both the op- tainment applications. Disney and Sega are example com- erator and the entertainment application providers despite panies. low prices per transaction. Revenue sharing for WAP games is based on airtime with regard to CSD and other Application Developer (AD) payment options for packet based connections. However, Application developers are responsible for creating en- the revenue for each depends on the revenue split among tertainment products. Codetoys is an example. the companies of the value chain. Difficulty securing favorable agreements with operators is one of the most fun- Publishers/Aggregators (Pub) damental limitations for mobile entertainment industry. Gartner analyst Ben Wood believes that revenue sharing Publishers fund the development of applications and is untenable as a primary revenue stream, with splits yield- bundle the applications from developers. ing no more than 10% for the content provider. [2] A fa- Publishers ensure a strong wholesale market position. vorable structure of NTT DoCoMo’s i-mode relations with Example companies include Digital Bridges and application providers is worth mentioning. They take a JAMDAT. 9% commission and pass the remaining revenue along the service chain. The success of NTT DoCoMo’s model Wireless Application Service Provider(SP) and much higher entertainment services usage of i-mode subscribers exemplify the value of mobile entertainment. Wireless application service providers develop, implement Therefore, it is important to establish favorable revenue or operate mobile entertainment platforms. Example com- sharing agreements among companies along the service panies include Cash-U, OpenMobile. chain to ensure business success. Portal Provider (PP) 3.2.3 Licensing Portal Providers provide the network based customer interface and the content selections through which a con- Some mobile entertainment companies may establish rev- sumer can access mobile entertainment products. Com- enue sharing agreements by licensing their services. In panies include wireless network operators such as this case, the payment structure changes radically to a Vodafone and Telefonica as well as independent portals series of transactions that occur prior rather than after such as Yahoo! Mobile. games deployment. Operators pay wholesalers for licensing their deployment technologies and applications, whole- Mobile Operators (Op) salers pay publishers for distribution rights to copyrighted Mobile operators provide mobile connectivity and con- applications, and publishers buy the copyrights from de- tent delivery between content repositories and consumer velopers. handsets. Vodafone, Telefonica, and NTT DoCoMo are examples. 3.2.4 Retail

Mobile Device Manufacturer (DM) In this model, users subscribe to the entertainment services with certain fees. To maximize volume, different Mobile device manufacturers are responsible for design, payments schemes can be used to allow frequently ac- manufacture and marketing of mobile devices and their cessed users to pay a flat fee for unlimited use and pay operating systems. Companies include such as Nokia, per use for casual users. However, at present only opera-

92 tors can employ this model, since they control billing for ranking would be offered prizes. The prizes include award- all end-user wireless services. Open mobile Internet mod- ing “points” to additional game play, merchandise at popu- els and secure m-payment technologies may eventually lar stores or certain e-commerce sites, gifts, holidays, and create opportunities for more direct and universal distri- cash. For example, Vodafone UK awards a cheque for bution channels involving independent portals. £3000 to top scorer entered the Hall of Fame who play- ing the popular mobile game “who wants to be a million- aire”. This provides a clever incentive for attracting and 3.3 Marketing Strategies retaining users. In order to successfully attract and retain the consumers requires effective marketing strategies. Registration and Subscription Registration and subscription services facilitate user data Co-branding Content Partnership collection and analysis, which in turn can be utilized for Establishing co-branding content partnerships with exist- user-based marketing strategies and service improvement. ing brands is the most effective way to reach a broad user User and applications databases can yield important de- base. With co-branding, the mobile entertainment com- mographic and play pattern data. Knowing who is using pany enters markets with the strength of wireless plat- which services and for what average length of time is ex- form while launching in conjunction with TV shows, tremely useful in evolving marketing strategies and con- movies, sports events, etc. The co-branding partners can tents design innovations. This would create potential op- utilize the strength of the company’s platform while lo- portunity of increasing profitability. However, securing calizing content and branding. Establishing such partner- customers privacy and without exposure to public is im- ships quickly will allow strong brand identity develop- portant. ment while the market is still in its infancy. [2] Marketing is a crucial aspect of success in a mobile entertainment venture. A successful strategy will utilize many Public Media of the above approaches. Establishing co-branding and It is well known that media formats such as television, partnerships is essential for maximizing exposure. This radio, and print provide significant opportunity to market will reduce marketing and advertising costs while pro- any products or services include mobile entertainment. viding association with a known brand. Prizes are an im- Press releases, advertisements, and wireless technology portant and effective strategy of successful mobile enter- publications relevant to target market segments can also tainment service. broaden exposure to a wide user base.

Online Advertising 4 CONCLUSIONS The primary consideration in determining advertising The mobile entertainment industry, while still in its for- strategy is how to most effectively reach the intended au- mative stages, shows signs of incredible potential. As the dience. Online advertising is now more interactive, visu- industry is young, complex and dynamic, there are doz- ally engaging, and personally relevant, this provides much ens of factors that will influence the way that the industry greater flexibility in content presentation. There are nu- develops. Some factors are interrelated, and there are some merous online advertising techniques. Employing rich remain unknown yet. media techniques to implement interactivity will improve This report outlines and analyzes the relevant issues and the overall effectiveness of online advertising. The mo- strategies for maximizing consumer demand and revenue bile entertainment has a unique advantage in terms of from mobile entertainment. Developing solid revenue advertising online branding due to its inherent interactivity. streams and highly competitive products is critical to re- Online demos can help gaining the most relevant possible alize the market’s full potential. This is a particularly com- exposure to the overall Web audience. plex market due to inherent technical restrictions, frag- mentation of the value chain between content creators and Promotions service providers, and the need for effective marketing to Promotions are also important and effective strategies for stimulate consumer demand. mobile entertainment companies or operators increasing Wireless technologies evolution is enabling increasingly sales and customer loyalty on the basis of incentive. In- diverse data communications and services possibilities. centives are an important factor of attracting consumers’ However, the pace of technological change and relatively stickiness. Consumers may be offered free logo, ring tone, deployment environments fundamentally limit the qual- and games etc. This is also an effective way of online ity and potential revenue for value-added wireless ser- branding. Contests and prizes are promotional methods vices. It is therefore necessary to optimize the services to of particular significance. Competition itself can some- meet customer needs according to commercial value and how entertain participants and prizes are the most appro- technology evolution. The complexity of relationships priate complement. Actually, most popular online games between entities in the value chain also needs to be ad- sites provide consumer top scores and overall rankings dressed to ensure maximizing profitability from mobile for each game, and the one positioned on the top of the entertainment for each role.

93 5 REFERENCES [1] Booz, Allen, Hamilton, Future Mobile Entertainment Sce- http://www.mobileentertainmentforum.org/articles/ narios, March 2003, Mobile Entertainment Analyst The_state_of_mobile.doc http://www.mobileentertainmentforum.org/articles/ [6] Today’s GSM platform, GSM Association, http:// MEFWP_Future_ME_Scenarios.pdf www.gsmworldcom/technology/gsm.shtml [2] Jennifer James, Mobile Gaming, August 2001 [7] WAP Forum, Wireless Application Protocol White paper, June 2000, http://www.wapforum.org/what/ [3] Simon Buckingham, SMS Tech, July 2001 WAP_white_pages.pdf [4] Simon Buckingham, SMS Express, July 2001 [8] WAP Forum, http://www.wapforum.org [5] Stephen N Wiener, The State of Mobile Entertainment,