Breakthrough Ideas (continued)
A Strategic Platform For Global Preemption
Atop its foundation of strategic architectural control of spread spectrum and a set of winning strategies in the standards war, Qualcomm tracks the dynamic path that sets the platform standards that shape, model, and mold its eventual global dominance of third-generation mobile telecommunications.
This claim contrasts sharply with a common assumption that market dominance of UMTS is preserved, even assured, in third-generation technology because the GSM oligopoly dominated the worldwide market of second-generation mobile wireless and built an outsized installed base. However, this competing claim counts its chickens before a European UMTS chipset is even hatched.
In the standards war over color TV, the FCC adopted the CBS mechanico-electrical system in October 1950. But, the market subsequently determined a different outcome. The CBS technology was not backward compatible, market-tested, nor ready for commercialization. Its unattractive price/performance ratio could not attract the necessary complementors. David Sarnoff of NBC snatched victory from the jaws of defeat by taking risks and redoubling efforts to develop NBC’s electronic version. Shapiro and Varian (1999, p. 217) noted, “…the color TV examples shows the danger of sitting back and assuming that you can maintain market dominance just because you control the current generation of technology or have a large installed base.”
In spite of the similarities of Qualcomm’s CDMA as the advanced technology to NBC’s electronic version, or of UMTS’s lack of compatibility with GSM, its absence of market-testing, and its delayed commercialization to CBS’s similar problems, no two sets of historical events are completely analogous. For instance, unlike the ETSI-announced choice of spread spectrum for its 3G system at the beginning of 1998, the color TV battle did not begin with CBS admitting that the future belonged to NBC’s electronic color TV.
Still, the premise of expected long-term economies in the scaling of R&D and manufacturing remains the fundamental rationale underlying any operator’s choice of GSM/GPRS/EDGE/UMTS as a 3G-migration path. This expectation of a 70 to 80% market share, which is derived by extending 2G’s present trends, remains a dubious forecast of an uncertain future. Such “forecasts” are nothing more than expectations-management-as-usual. Just as it would not be meaningful to project the present trends in rollouts by operators of cdma2000 1X and UMTS, where CDMA is ahead by 18 to 1, or by the present number of subscribers, where CDMA is ahead by 100 to 1, as predictive of the future.
Given that such “trend-forecasts” are meaningless, are other claims of GSM-UMTS leadership more reasonable? How do the economies of scale in R&D for the GSM/GPRS/EDGE air-interface transfer to knowledge of how specifically to commercialize the spread spectrum interface? How does experience gained with manufacturing GSM chipsets trim down the design-learning curve for the switch to a new RF system for modulating spread spectrum? Do such forecasts consider Qualcomm’s three breakthrough ideas that shape the future?
Ordinarily, the value of an installed base derives from the combined switching costs of users because it is so difficult to coordinate a switch to a new standard. That is, the aggregated switching costs become non-linear due to the coordination problem. But, it only takes one operator, not all of its customers, to choose to switch to GSM1x. In this instance, because 2G GSM/TDMA networks must switch to spread spectrum as the mandated air-link standard for 3G networks, switching costs for operators are simply unavoidable. Instead of paying later, GSM1x lets you pay now to earn more now. After one carrier operates a GSM1x network, then demonstrated excellence will prove excellence.
Whereas Europe closed down most of its analog systems to require the move to GSM, the purchasers of 3G spectrum for UMTS do not own franchises that cover identical territories held by the GSM-operators. Most pundits foresee UMTS rolling out in urban areas, with GSM remaining common in rural areas. In Europe, delays follow delays that follow pushed-back estimates of rollouts. Even if you set aside the one billion Euros spent on new spectrum, the longer, delayed, and more costly 3-billion-Euro path for GSM/TDMA migration to UMTS, in fact, is a misfired and failed political strategy that handicaps both its operators and their customers.
Customers remain so unaware of standards that churn is common because all that is required to switch is the cost of a new handset and paying for the remaining time on a contract. Customers seek advantages in value based on useful and meaningful functionality. Their lock-in, as Shapiro and Varian indicated, is not so costly.
The lock-in for Europe, however, is quite costly, perhaps insurmountable because the politicians may not ever let the operator’s choose any wireless system that is competitive with UMTS and not already mandated. But what Europe denies to itself, it cannot deny in the rest of the world.
This longer UMTS migration path becomes more costly because it requires new handsets and extra base stations, but also, it handicaps operators even more because it offers lower than promised data rates with GPRS and EDGE that steal capacity from already capacity-constrained voice. Not only does costly migration fail to deliver the superior performance and capacity of cdma2000 1X that is available now, but also after adding the GPRS and EDGE sunk costs, the entire air-link for the UMTS-version of spread spectrum still requires a swap-out.
The contrast with cdma2000 is stark. Given its compensatory increase in voice capacity, higher data rates, and predictable evolution to higher levels of performance, the cdma2000 GSM1x alternative becomes attractive, particularly as the full suite of radioOne architecture becomes available in 2003 to harmonize all 2G and 3G modes and bands into one seamless network.
From the battle over DC/AC electrical systems, one lesson learned was that a first-mover advantage could be overcome by a superior technology if the performance advantage was sufficient and users were not overly entrenched. The performance advantage appears to be sufficient, but, as 3G UMTS loses GSM’s 2G leadership around the world, the European Union will strive to keep its entrenched base in Europe intact. This weary European socialism grows old, handicapping its companies in a globally competitive entrepreneurial era. But also, it is understandable because Europe lost its competitive positions in consumer electronics and computers. Wireless telephony has been its entrepreneurial pride and joy.
However, by protecting its workers and its companies, Europe fails to accept that principles of economics mandate that job and industry protectionism must fail because the Darwinian price of progress requires that people must adapt by leaving old jobs for newly created entrepreneurial opportunities in new niches. If it comes down to survival of the fittest, then the sustainable competitive advantage of strategic architectural control will select who is fit.
Qualcomm’s Strategic Control of an Expanding Platform of Architectures
At the May 22 2002 Analyst’s Conference, COO Tony Thornley presented the mission [objectives] of Qualcomm’s global business: (1) to expand CDMA worldwide, including cdma2000 1X, 1xEV-DO, WCDMA, and radioOne; (2) to drive replacement cycles using a consumer-electronics model for differentiating new products; (3) to add new revenue streams from BREW, QChat, and Wireless Knowledge’s enterprise applications; and (4) to expedite the development of new areas of knowledge, including Digital Media and, following 9/11, homeland security. To deliver these objectives, Thornley reviewed [with many additions from Don Schrock’s slides] Qualcomm’s set of enabling building blocks for building technologies and capabilities.
The Basic ASSP Platform.
Commercially available today, the 8th-generation MSM5xxx chipsets excelled by introducing five different CDMA technologies in one generative year, 2001: (1) the MSM5100, which was based on the MSM3300, enables cdma2000 1X Rel A, with its peak data rate of 307 kbps on the forward link, and includes gpsOne position location, and standards such as Bluetooth, USB, MP3; MIDI ringer/CMX, MMC, and U-RIM; (2) the MSM5500, which was based on the MSM5100, enables 1xEV-DO, with a peak data rate 2.4 Mbps on the 1xEV-DO forward link, and includes gpsOne position location, and standards such as Bluetooth, USB, MP3; MIDI ringer/CMX, MMC, and U-RIM; (3) the MSM5200 enables UMTS, with a peak data rate of 384 kbps, and includes a new memory interface, U-SIM, and MP3; (4) SnapTrack’s gpsOne enables location-based services; and, (5) BREW, the consumer and enterprise applications platform, enables new and useful wireless data applications. At present, the MSM5100 is the largest seller.
Of course, Qualcomm provides compatible Cell Station Modem (CSMs) for base stations that support and correspond to the features of the MSMs, specifically the CMS5000 for 1X networks and the CSM5500 for 1xEV-DO for high data rate. These solutions facilitate higher numbers of users per channel and backward compatibility to IS-95. They can be easily added as line cards to increase network capacity or to support higher-data-rate services. The CSMs come with a full set of documentation and system API software to reduce development time and cost, encourage consistency in implementing all CDMA infrastructures, and leverage existing CSM expertise and design knowledge. To date, no other company has introduced competitive CSMs into the IS-95 or CDMA2000 market, ceding 100% of the market to Qualcomm (NEC provided DoCoMo’s cell site modem for FOMA’s WCDMA).
The DMSS dual-mode (CDMA/AMPS) subscriber software provides a complete software solution for building wireless products, including complete documentation, applications engineering support, and intensive training. DMSS runs on QCT’s SURF development platform, which provides assistance in evaluating, testing, or debugging software. QCT has a complete CDMA Designer Toolkit that includes support for CDMA chipsets, system software, ARM microprocessor, QDSP processor, and CDMA analysis and test tools, including an easily reconfigurable Subscriber Unit Reference (SURF) platform and other tools designed to allow manufacturers to shorten design time, increase quality, reduce costs, and differentiate their own solutions.
Sampling now and commercially available in 2003, the 9th-generation MSM6xxx family of chips definitively establishes product leadership in spread spectrum platforms because: (1) its superior system performance, broadest feature sets, and radioOne architecture represent best-in-class products; (2) its 50% reduction in RF board area, its reduced assembly cost, and 30% reduction in the bill of materials produce lower cost solutions; (3) its complete solution that uses the same interfaces and system architecture for all modes reduces time to market; and (4) its chipset architecture facilitates seamless multimode/multiband design that enables new markets. Moreover, as discussed in Part II, it includes Qualcomm’s new Selectable Mode Vocoder that increases channel capacity significantly while maintaining excellent voice quality. Unlike GSM’s AMR vocoder, only Qualcomm’s SMV immediately translates any reduction in average vocoder rate into a corresponding capacity increase over present 1X capacity.
Qualcomm’s third breakthrough idea was commercialized as the revolutionary radioOne system solution based on new software, RF-, and digital-baseband technologies. This new architecture simplified the chipset by removing the need for Intermediate Frequency stepups-or-stepdowns to UHF and, instead, directly converts RF-to-baseband or baseband-to-RF signals. The analog voice signal of the MSM6xxx and the UHF signal of the antennae directly converse through the translational mediation of the RFR6000 in the handset. The radioOne architecture, in the 6xxx family is accompanied by MMSS (MultiMode Subscriber Software), which is comparable in functions to the DMSS, but that integrates functions across all modes and bands.
The ZIF-simplification and radioOne’s breakthrough multimode integrations provided Qualcomm with a second (in addition to its high data rate lead enabled by 1xEV-DO) multi-year lead in product development. The Qualcomm recipe for combining these multiple breakthroughs redefined the nature of the game in 3G mobile telecommunications by introducing revolutionary architectural advances within the Qualcomm platform and strikingly lowered the cost of the handset BOM by 30%.
This ZIF-simplified multimode architecture allowed Qualcomm to segment the mobile terminal market into three tiers. Specifically, the MSM6000 is designed as an entry level, low-tier, low-cost phone, which is based on 1X Rel 0 and optimized for voice, but permits SMS through it 14.4 kbps data rate. The MSM6050 is designed as a low-end terminal, which is based on 1X Rel 0, and provides optimized gpsOne, 153 kbps FL/RL, turbo code support, MP3, CMX, and MIDI support, JPEG encoder/decoder, low frame rate MPEG-4 decoder, and a new memory interface. Its MPEG-4 encoder permitted the first demonstration of mobile wireless video capability integrated on a chip.
Three full-featured chipsets were designed for the mid-tier market: the MSMS 6100, 6200, and 6300. Based on the MSM5100, the MSM6100 is a cdma2000 1X Rel A chipset and features radioOne architecture and an ARM926EJ-S CPU to provide a full set of multimedia capabilities. Thus, the MSM6100 provides 307 kbps FL/RL, assisted gpsOne, Java H/W accelerator, H/W graphics accelerator (2D/3D), enhanced liquid crystal display interface, MPEG-4 decoder/low complexity encoder/ fast JPEG encoder/decoder, CCD/ CMOS sensor camera RFs, a selectable mode vocoder, and a new memory interface.
Based on the MSM5200 and MSM6100, the MSM6200 enables peak data speeds of 384 kbps for UMTS in its multimode GSM/GPRS/UMTS handset, with much of the multimedia set of the 6100. [The MSM6250 has been added to this lineup.]
Based on the 6100, with its full multimedia, the MSM6300 for GSM1X is a multi-mode chipset for cdma2000 1X Rel A and GSM/GPRS. If you view slide #122 from the Analysts Conference, you can see a graphic of the 6300’s architectural layout. The solution consists of four integrated chips: MSM6300, RTR6300 (transceiver), RFR6000 (receiver), and PM6050 (power management). The RTR6300 is a fully integrated GMS/GPRS transceiver, with the GSM/GPRS processor on the MSM6300, along with the CDMA processor, gpsOne processor, advanced ARM processor, QDSP4000 processor, Bluetooth processor, controllers, ADCs, interfaces, and the like. Thus, various coprocessors become the building blocks of multimode functionality, which are all tightly integrated into a single circuit.
Two high-tier multimode MSMs will sample in the second half of 2002, becoming commercially available during 2003. Based on the MSM6100, the MSM6500 integrates the functioning of cdma2000 1X Rel A and GSM/GPRS/UMTS into a single integrated MSM to create a world phone. Based on the MSM6100 and MSM6200, the MSM6600 integrates cdma2000 1X Rel A, 1xEv-DO, and GSM/GPRS/UMTS into a single MSM to create an all-purpose high data rate world-phone.
In 2004-2005, the MSM 7xxx family of chipsets will enable more advanced applications and improved user interfaces. The 7xxx family will increase the scope and functionality of multimedia, increase the functionality of speech commands and other input devices, and permit the continuing advanced performance and speciation of output devices to fill new niches. It completes the transition to receiver antennae diversity that doubles voice capacity for a second time within the 1X architecture.
In effect, each new family evolved from its parents to increase functionality along sustained trajectories of performance, extend scope, and fully use its increased processing capacity to introduce more complex advanced coding that, in turn, drives further evolutionary advance. The key goal of advanced performance is to facilitate the use of productive and entertaining applications that drive ARPUs because they are valuable, easy-to-use, and meet the needs and priorities of operators, manufacturers, and end-users.
All of these products are scheduled to appear before any future UMTS networks grow to a significant total size, say, 10M subscribers, perhaps in 2005. In contrast to continuing GSM/GPRS/UMTS delays and technical disappointments, through June 2002, Qualcomm had sold 31M cdma2000 1x chips. These achievements in the first half of 2002 took place on top of a difficult 2001 for chipmakers. A year, however, in which Qualcomm proved to be the exception in the chip industry by out performing to the upside the overall downside business trend in semiconductors. In fact, the company attained the #1 share of sales in Application Specific Standard Products (ASSPs) in 2001 by registering about a 10% increase in its chip sales to 1.4B dollars.
Qualcomm’s basic ASSP platform provides seamless scalability from entry-level to high-end mobile terminals because its integrated learning base in spread-spectrum and chip design provides a basic platform that can be translated into multiple tiers of products that are differentiated by increased scope of performance. Driven by new features and services, devices are segmenting into phones at different prices and levels of data usage, ranging from inexpensive low-data-use voice/SMS phones to more expensive video conferencing devices at the highest end of cost and usage. The lower tier single-use devices increase data usage as they move from voice/SMS to meter readers, browser phones, and on to smart phones and music phones. In the mid-tier lineup, increases in data usage flow from location trackers, data cards, camera phones, and PDA/phones, to large-user-interface web data devices. The higher priced tier increases data usage as its devices move from navigation to game phones and small-form factor specialty phones (in, say, sunglasses) to devices specifically designed for telematics, video streaming, and video conferencing. Gartner Group estimated that the total market is segmented into 42% basic handsets, 56% advanced handsets, and 2% Smartphones (phone/PDA combination). In addition to this complete segmentation of devices, Qualcomm has the largest base of 3G OEM customers, some 50 +, with over 120 1X/1xEV-DO devices in design or manufacture.
Because it continued to use the same protocol stacks, same unique RF encoding to identify users, and same method of spreading the spectrum, Qualcomm’s 3G air-link interface provides both backward compatibility and future stability to this evolving and expanding platform.
The basic ASSP platform consists of a set of architectures organized as a single integrated amalgam, a tightly integrated circuit consisting of coprocessors that perform specific processing functions for CDMA, high data rate, position location, multimedia, DSP, Bluetooth, or GSM/UMTS, while using ARM’s core for CPU processing.
Qualcomm knows how to introduce new features into each coprocessor building block to take advantage of advances in Moore’s Law, RF modulation, diversity technology, newly introduced or upgraded connection standards, and advanced methods of coding.
Like Intel, its design teams can simultaneously integrate their specialized, improved, functional algorithms into continually overlapping sets of concurrently developing and functionally overlapping future generations of compatible chipsets. Unlike Intel’s in CPUs, Qualcomm’s design decisions have not become locked in by the need for backward compatibility to less elegant earlier architectural choices because the architecture for its air interface remained stable, ensuring effortless backward compatibility. Thus, the air interface evolved to increase its unremitting performance advance along sustained trajectories of performance while adding new multimode and multimedia building blocks to expand functionality.
The single integrated chipset became a basic ASSP platform hat stabilizes interfaces, conserves energy, reduces size, and speeds both time-to-market and performance. Also, the basic platform began to expand by adding both multiple modes and bands that harmonize systems and multimedia functionality suited to downloading newly developed applications using the BREW API.
When compared to a PC, issues of optimal size and power in a mobile terminal represent major constraints that limit functionality in mobile wireless. Hence, Qualcomm’s engineering maxims to reduce size, conserve power, and provide complete standardized solutions. Given the improving price/performance ratio generated by Moore’s law and because advances in functionality come from increasing the complexity of coding to take advantage of new knowledge, there is a continual confrontation of design skill versus resource constraint. Qualcomm’s talented design teams coalesce scientific comprehension with practical expertise: how to code efficiently to best use increases in processing power or advances in RF modulation and simultaneously how to do so while conserving energy and reducing size. The improvements in performance and the expansion of functionality necessitate integration that microscopes ever more elegant and sophisticated coding into a single integrated circuit using less board space. This is Dr. Gelernter’s (1998) Machine Beauty, the elegance at heart of technology, and the single most important element in design, beauty resulting from an inspired mating of power with simplicity.
By substituting spread spectrum coding originally, and either increased simplicity or advances in processing-power continually, Qualcomm mastered the scale-downs that reduced size and transmission-power. Once performance solutions were found, time, money, and energy were put to tight integration that reduced transmission power further. Then, its scientists and engineers mastered the architectural simplifications and tight engineering that scaled down board-size. Its deep-craft mastery magnified beautiful, breakthrough ideas that Qualcomm codified-as knowledge and commercially converted into accelerating economic value. For instance, its remarkably innovative radioOne architecture in the family of MSM6xxx chipsets replaced the heterodyne approach using intermediate frequencies that had been introduced way back in 1917. This was a rare revolutionary triumph in architectural reduction.
This triumphant simplicity proved to be a structural necessity because it permitted the radioOne technology’s small form factor that enables the coming global harmonization of all third-generation architectures, including not only Qualcomm’s unique solutions for universal frequency reuse and for optimizing high data rates but also its rivals’ less elegant architectures. The future textbooks in business will analyze and laud this case of extraordinary success in a knowledge company. |
