Shosteck Group White Paper on TDMA 3G Migration Paths  |
|||||||||||||||||
 |
|||||||||||||||||
|
An analysis by GSM or CDMA: The Commercial and Technology Challenges for TDMA Operators Also Available in Portuguese 1 An Overview 1.1 Introduction This white paper reviews and discusses the options open to TDMA/IS-136 network operators for migrating to third generation (3G) wireless technologies. In particular, it addresses the suitability of CDMA2000* 1X (formerly called CDMA/IS-95-C, CDMA 1xRTT, or cdma2000 1x), and its future derivatives CDMA 1xEV-DO and CDMA 1xEV-DV, as a 3G alternative. Central to our analysis are the issues of handset availability, handset cost, and economies of manufacturing scale over the short- to medium-term. In evaluating them, we find that in transitioning from TDMA/IS-136 to 3G, the advantage often attributed to mature technologies may not hold. The Shosteck Group does not endorse one technology over another or one 3G-migration path over another. Rather, we point to and discuss possible challenges that TDMA/IS-136 operators may face in deploying GSM and possible advantages that cdmaOne™† may offer. Based on these possible advantages, we suggest that TDMA operators may find it worthwhile to evaluate CDMA as a 3G option. Among TDMA operators who do so, some may conclude that GSM remains their preferred choice. This may be especially so for those licensed in 1900 MHz frequencies. Other TDMA operators may conclude that CDMA provides a better choice. This may be especially so for those licensed in 800 MHz frequencies. We intend this white paper to raise core issues concerning the 3G transition. The resolution of those issues, to the extent feasible, must take place in discussions between the network operators and their vendors. As a first step, we clarify what we mean by 3G. The radio component of advanced technologies varies in terms of at least three characteristics.
The International Telecommunication Union (ITU) serves as the arbiter of 3G standards. It does not define 3G in terms of channel width or spectrum allocation, but rather data rates. By ITU definition, the 3G RF interface can deliver data rates of 144 kbps or greater. The ITU recognizes W-CDMA (UMTS) and CDMA2000 1X as meeting this criterion.1 Higher data rates will enable end-users to experience richer content than is now available and, in conjunction with packet architecture, to gain instant and low cost access to the Internet. Higher data rates, and especially instant and low cost access to the Internet, will expand future network traffic. As network traffic expands, operator revenues will increase. However, independent of the capability of technologies, operators must recognize the economic-commercial trade-off of network costs versus data rates. The higher the data rates, the greater the network costs. Eventually, every operator must optimize the data rate it offers to end-users in terms of the cost to provide it versus the revenues it generates. Currently, TDMA/IS-136 operators AT&T and Rogers AT&T have chosen to adopt GSM, from there to progress to GSM-GPRS, then to deploy GSM-GPRS-EDGE, and finally to adopt UMTS (also referred to as W-CDMA). This last step takes for granted that spectrum for UMTS will be available. This currently assumed migration path has evolved rather recently. Other TDMA operators are still deciding which migration path to pursue. A year ago, the assumed migration path for TDMA/IS-136 operators was to deploy TDMA-GPRS, then TDMA-GPRS-EDGE, and afterwards—through an ill-defined process—to adopt UMTS. This earlier migration path is no longer being discussed. In reviewing the migration paths, we focus on the commercial issues with which TDMA/IS-136 operators must deal in choosing a 3G alternative. Five sets of issues stand out. These are:
These latter two are especially important. If operators cannot deploy viable commercial services in a timely manner, they run the risk of losing competitive position. If the network and/or handsets deliver a negative end-user experience, they will not only fail to generate revenue, but will encourage churn as well. We pay little attention to the nuances of technology legance that alternative technologies may or may not provide. For our purposes, the commercial issues are overriding. Not to be forgotten is the importance of voice. Using compression techniques and more sophisticated bandwidth management, 3G will eventually enable non-voice applications such as full-motion video and multi-media, all in real time. That said, voice will continue to provide operators with the bulk of their revenues into the foreseeable future. This stems from the “data stimulus effect.” Increased data use does not displace voice traffic but expands it. Even the most enthusiastic supporters of non-voice applications recognize this phenomenon.2 1.2 Migration Paths and Time Frames for Deployment In 10 to 15 years, the current issues of 3G migration will be dimming into historical memory. All of today’s 2G operators will have deployed 3G and, possibly, yet more advanced technologies. The deployment challenges, some not yet fully recognized, will have been successfully surmounted. Today, the critical issues that operators face center on what next generation technology paths they choose for the immediate- to mid-term future—the next one to five years. For operators who now deploy GSM, cdmaOne, or PDC, the one to five year evolutionary paths are clear. They are much less so for operators who deploy TDMA/IS-136. To review, GSM operators, or more precisely those assigned 900 and/or 1800 MHz spectrum, will evolve first to GSM-GPRS and eventually to UMTS. They may or may not deploy GSM-GPRS-EDGE as an intermediary step. Deploying UMTS will require these operators to use newly allocated and assigned 3G (UMTS) spectrum at 1900 (uplink) and 2100 MHz (downlink) in conjunction with their currently assigned GSM spectrum at 900 and/or 1800 MHz. It will also require multi-mode/multi-band GSM-GPRS-UMTS (or GSM-GPRS-EDGE-UMTS) handsets. Such multi-mode handsets will enable handoff from network to network. This will enable seamless provision of basic GSM services (voice and messaging) throughout the network and provision of UMTS in the heaviest traffic parts. This also enables operators to deploy 3G infrastructure only as demand requires, thus minimizing their investment burdens. The migration path for GSM operators assigned 1900 MHz spectrum (almost all in the Americas) is less clear. Separate 3G spectrum, suitable for UMTS, has yet to be allocated. Until such spectrum is allocated and cleared, this precludes GSM operators at assigned 1900 MHz spectrum from advancing beyond GPRS and EDGE, if the latter becomes commercially available. cdmaOne operators, whether assigned frequencies at 800 MHz, 1900 MHz, or both, can evolve to CDMA2000 1X using their current spectrum. This obviates the challenge of finding new spectrum. The evolution to CDMA2000 1X requires channel cards and software upgrades to cdmaOne base stations and introduction of CDMA2000 1X handsets. All CDMA2000 1X handsets are backward compatible with legacy cdmaOne infrastructure. This obviates the need for multi-mode handsets. In Japan, PDC operators will construct entirely separate UMTS networks on newly assigned 1900 and 2100 MHz spectrum. Subscribers will access those networks with single-mode/single-band UMTS handsets. During the next one to five years—or through mid-2006—no operator, regardless of whether they are today GSM, cdmaOne, TDMA/IS-136, or PDC, will complete the evolution to 3G. Some may fully deploy 3G infrastructure. However, sans massive handset subsidies, all will still have subscribers who continue to use the 2G network. Some operators will advance substantially in making the transition to 3G. This will be particularly so for those who currently deploy cdmaOne and adopt CDMA2000 1X. For those who currently deploy TDMA, the evolutionary path is more complex and less clear. That said, regardless of whether TDMA/IS-136 operators choose GSM or cdmaOne as their migration path, they will have to overcome challenges, which the industry has never before faced. Because of this, both paths will prove more complex, expensive, difficult, and time consuming than many initially imagined. However, as we suggest in our following analysis, for at least some TDMA operators, CDMA may prove the less onerous alternative. 1.3 The Dilemma of TDMA Operators The greater complexity, expense, difficulty, and time will stem from TDMA/IS-136’s becoming an “orphan technology.” Orphan technologies are characterized by twolimitations.
Such orphans may be new technologies with only a limited life cycle. High rate circuit-switched data (HSCSD), which was barely deployed before being abandoned, provides an example. They may be mature technologies at the end of their life cycles. TDMA falls into this category. By their inherent nature, orphan technologies are being superceded by more functional and/or lower cost alternatives. For this reason, they provide no more than short-lived advantages for network operators and/or end-users. For manufacturers, this means quickly diminishing production volumes and profits. As manufacturers recognize this, they stop R&D investment. This accelerates the orphan status of the technology. With that, it falls more behind newer and more advanced ones. As it falls behind, it provides fewer and fewer benefits for network operators and/or end-users. Eventually, it is abandoned. Other examples of orphan technologies include the telegraph (superceded by the telephone), telex (superceded by fax, which, in turn, is being superceded by email), and circuit-switched telephony (being superceded by packet telephony). Worldwide terminal sales attest to the emerging orphan status of TDMA/IS-136. During 2000, 63 percent of terminals sold were GSM, 13 percent were cdmaOne, and nine percent were TDMA.4 The relatively small volume for TDMA has made it the least attractive for manufacturers. The orphaning of TDMA was sealed by the announcement of AT&T Wireless, made in November 2000, that it was abandoning plans to incorporate GPRS and EDGE into TDMA. Instead, AT&T announced that it will overlay GSM onto its current TDMA network.5 In concept, this will enable AT&T to deploy GSM-GPRS, GSM-GPRS-EDGE, and, assuming available spectrum, GSM-GPRS-EDGE-UMTS as its migration path to 3G. AT&T’s approach is premised on AT&T’s taking advantage of the massive R&D efforts that are being devoted to the GSM transition path (i.e., GSM-GPRS and GSM-GPRS- EDGE) and the expected economies of manufacturing scale of GSM, and eventually of UMTS, over time. Indeed, this premise of long-term economies of R&D and manufacturing scale is the fundamental rationale underlying the choice of GSM as a 3G-transition path.6 It defines the central, albeit unspoken, dilemma that each TDMA operator faces. On the one hand, should the operator choose a 3G path with lower short- to mid-term costs but uncertain long-term costs? Or on the other hand, should the operator choose a 3G path with higher short- to mid-term costs but apparently more certain and seemingly lower long-term costs? A goal of this paper is to clarify these cost issues. Later, we discuss the economies of manufacturing scale. We explain that by deploying dual-mode TDMA-GSM handsets, TDMA/IS-136 operators will be unable to reap the economies of scale commonly associated with adopting GSM as the path to 3G. We also discuss the trade-off between cost-efficient (and potentially profitable) network design and operation and the provision of high data rates. In adopting GSM, AT&T ended any hope for further development of TDMA/IS-136 and a migration to 3G based on it. This means that all TDMA operators must now choose a new migration alternative. While most industry observers have assumed that this would be GSM, it could be cdmaOne as well. Using GSM as a means of reaching 3G is an appealing concept. However, it requires that TDMA/IS-136 operators deploy what, in effect, will be a completely separate 3G network. This will prove a challenging and expensive proposition. It will be especially challenging to TDMA operators that are assigned to the 800 MHz rather than to the 1900 MHz frequencies. For the 800 MHz operators, the availability of GSM infrastructure is still a promise, although Nokia, Ericsson, Motorola, and Nortel, who have made such promises, are credible vendors. Far more important, and still unspoken, will be the timely delivery of 800 MHz GSM handsets at a reasonable cost. As of this writing (June 2001), both their date of delivery and cost remain unknown. 1.4 The Two Steps of 3G Evolution Regardless of whether TDMA/IS-136 operators choose GSM or cdmaOne as a migration path to 3G, they face the same two transition barriers.
That said, GSM and cdmaOne offer substantially different migration alternatives. We explore these in the chapters that follow, looking especially at the issues of timely deployment and cost. 
|
|||||||||||||||||
 |
 |
|
|
||||
|
||||
|
|