re: Vintage Grahame Lynch
<< I have just received GL's book.>>
Message 16066687
Message 15989034
Message 16436991
My comments on the above:
Message 16438526
Another classic from my archives (adapted from his book) that unfortunately is no longer available from America's Network archives:
>> How GSM Beat CDMA
Grahame Lynch
America's Network
June 1 , 2001
americasnetwork.com
Statistically, the biggest telecom phenomenon of recent years wasn’t the Internet — it was the GSM phone. And while GSM may be co-opting the CDMA radio interface for its next generation, this doesn’t mean victory for the CDMA forces, argues Grahame Lynch. This article is excerpted and adapted from his new book, “Bandwidth Bubble Bust,” available from online book stores this month.
Ask someone to name the biggest revolution in telecommunications over the past decade or so and many would reply “The Internet.” But on sheer numbers, they are wrong. The biggest revolution has been the cellular phone.
In the United States alone, about 67 million people use the Internet once a week or more. By contrast, 114 million use mobile phones.
Internationally, the comparisons are even more extreme. Across Western Europe, cellular penetration rates typically exceed 50% of the population. Less than one-third this number qualify as active Internet users.
For many, the Internet represents the triumph of the market. Despite its governmental beginnings, its growth has occurred outside the mandate of official encouragement and institutional endorsement.
Cellular is something quite different. Although few predicted its current market success, much of the development of cellular has occurred as a deliberate result of government planning.
Some 75% of all the digital cellular phones in the world use a standard that was conceived in Europe throughout the 1980s and commercialized in 1991. This standard, Global System for Mobiles (GSM), was very much designed with the interests of vendors and the political interests of European unification and industry policy at heart.
GSM uses a time-coded radio air interface that provides a good quality signal but suffers capacity limitations.
This forced operators to buy more infrastructure as their networks became loaded. But the standard compensates for such costs in other ways. It provides for standardized billing and signaling systems, opening the way for such nifty features as international roaming. Handset intelligence is based in a removable smart card, called a Subscriber Identity Module, which makes it easier for users to change handsets and customize their settings. GSM also incorporates the capacity for users to originate and receive short text messages of up to 160 characters in length.
To many American observers, GSM was a typical case of continental over-engineering.
In the US, the preferred standard was a time-based air interface (TDMA) that was simply grafted on to the existing signaling systems used by the so-called first-generation analog networks. Japan went down its own idiosyncratic path, developing the Personal Handyphone Service (PHS) and Pacific Digital Cellular (PDC) standards.
But over-engineered or not, GSM had an advantage. Not only had the Europeans developed the most fully featured standard on offer, but they also boasted the most globalized vendor channels.
Government spectrum managers worldwide rarely had the resources to conduct much of their own research on emerging technologies, so they usually relied on vendor advice. And there was no shortage of European vendor representatives ready to offer advice to license radio spectrum for GSM networks. This advice was quite clearly heeded.
By 2001, there were over 400 GSM networks operating in 162 countries across the world. Just two major economies — Japan and South Korea — had resisted the push.
The San Diego Juggernaut
With Europe and SE Asia embarking on the GSM path and the US pursuing the 1.5G technology of TDMA, there was an obvious opportunity for a major “disruption.”
And so it came, from San Diego firm Qualcomm and its CDMA technology.
While TDMA used a timing mechanism for its radio interface, CDMA used a coding mechanism. Qualcomm and its promoters claimed that this offered capacity benefits some tens of times that of the capabilities of existing analog and TDMA-based networks. In the minds of many, CDMA was the next big thing.
CDMA was nothing new. It was originally patented by none other than actress Hedy Lamarr, who, together with her husband, gained inspiration for the technique from observing the musical coding on piano rolls. From the 1950s on, CDMA developed into a staple US military technology.
Qualcomm’s real achievement was to take these ideas and apply them as a viable cellular technology. Driven by founders Dr. Irwin Jacobs and Andrew Viterbi (who are still involved with the company to this day), it demonstrated a system as early as 1989. Jacobs proved an effective propagandist for the technology. The ensuing technical debate between the CDMA and TDMA camps was quickly labeled the “cellular holy war.”
However, Qualcomm had no manufacturing experience and European vendors weren’t interested in funding its intellectual property rights. At first, Motorola proved the only vendor able to push the technology to commercial deployment stage.
While GSM systems were entering service across Europe, Asia and Australia throughout 1993, the first CDMA system wasn’t to make an appearance until two years later — in Hong Kong, and then a few weeks later in California.
But only a few operators in the Americas committed to using the system. Early reports were negative. Networks didn’t perform to expectation and handsets were in short supply. CDMA had a big voice, but few results to show for it.
The Korean Advantage
CDMA’s first big break occurred in South Korea in 1996. Throughout the early 1990s, the South Korean government had studied ways to develop a comparative advantage for its electronics manufacturing industry.
Correctly observing that European vendors had locked up the GSM market, the government decided that the country’s carriers should pursue the CDMA path instead. The American CDMA forces, hungry for export success, eagerly licensed them the technology.
By all accounts, Korea’s main wireless operator, SK Telecom, wasn’t too happy at being forced to deploy a CDMA solution supplied by a Korean vendor under US licensing. With the inevitable delays of such an undertaking, SK Telecom wasn’t able to launch a CDMA network until 1996.
The CDMA forces proved a little too guilty of believing their own propaganda. The early technical performance of CDMA was dreadful. One credible independent study showed it up as the worst of the three major standards. Conducted by a technical department of Hong Kong University in 1997, the study measured the performance of Hutchison’s CDMA network against Pacific Link’s TDMA network and some six GSM networks operating in the 900 MHz band and the newer, immature 1800 MHz band.
The CDMA Forces Took The Game To Another Level.
The CDMA network was easily the worst for call dropouts, losing as many as ten times many calls as the better-performing GSM networks. Vendors were slow to fix these problems, because they had little experience with the core IPR and Qualcomm was too removed from the problems to offer speedy solutions. A study conducted in Singapore in 2000 gave the same result — GSM performed much better than CDMA in the field.
The Qualcomm forces, bolstered by the incredibly hyperactive organizing efforts of Californian PR specialist Perry La Forge and his CDMA Development Group, pressed on with their propaganda claims throughout 1997 and 1998.
But as European vendors did better with their GSM standard, eventually cracking the Shangri La of mass adoption in China, the CDMA forces decided to take the game to another level.
The China Prize
The US State Department was conscripted to the cause, and the contemporary Secretary of State, Madeline Albright, led the charge. Letters were dispatched to Europe complaining of that continent’s restrictive trade practices in favoring GSM over CDMA.
But the real prize was China. The State Department realized that Chinese adoption of CDMA constituted an effective bargaining chip over that country’s admission to the World Trade Organization. Throughout 1998 and 1999, the CDMA/WTO trade-off became a key discussion point among China’s top leaders. American negotiators, including those representing the CDMA lobby, couldn’t help themselves, regularly discussing the progress of talks to the likes of the Asian Wall Street Journal. Those of us in the Chinese telecom reporting fraternity (I was one at the time) marveled at the naiveté of American negotiators who were flagging their tactics and agendas through press leaks.
The Chinese, living up to reputation, played a wonderful game of ambiguous negotiation. Premier Zhu Rongji would make positive noises about CDMA adoption. Then another top official, such as Communications Minister Wu Jichuan, would state something to contradict him.
As part of the push, the US State Department began pressuring allies in the Asian region to also adopt CDMA — a novel form of “cellular encirclement.” Several countries in SE Asia issued CDMA licenses to obscure and under-funded operators. At time of writing few had launched, and one in Singapore had closed down. Taiwan’s government-owned Chunghwa Telecom also announced a CDMA network, only to quietly withdraw its plans some time later.
The US strategy was revealed in Australia in 1998 when the Australian government issued a press release stating that Telstra would build a CDMA network one day before Telstra’s board was actually due to meet to consider its own decision. My Australian newsletter, Communications Day, reported at the time that Telstra had been offered a “virtually free” TDMA network from Ericsson. It later emerged that the US State Department had lobbied the Australian Prime Minister to adopt CDMA.
The 2,600% Year
The second big break for Qualcomm came as standards bodies across the world came together to decide on the replacement for GSM and TDMA.
Almost everyone agreed that CDMA provided the best air interface for the high-capacity needs of future services. Although there were many competing proposals vying to get acceptance from the International Telecommunication Union, almost all of them were based on CDMA. Qualcomm was adamant. All of them would have to pay royalties to Qualcomm. It seemed that Qualcomm would get its eight cents worth of every dollar generated by cellular equipment into the distant future.
CDMA also seemed like it was on the verge of conquering North America. A number of fast-growing networks operated by the likes of Airtouch, Bell Atlantic and Sprint were using the technology. CDMA networks had also been licensed in the virgin markets of Latin America. Meanwhile, the Korean networks were setting growth records.
Combined with the expectation that China would adopt CDMA as a condition of WTO entry, Qualcomm’s stock price soared. It soon became one of the most widely held stocks in the United States. The stock’s value rose an incredible 26 times in 1999, ending the year at a value of $200 per share and a market capitalization of over $140 billion.
Subsequent announcements fuelled the juggernaut. Korean and Japanese CDMA operators appeared set to upgrade their networks to the next-generation cdma2000 standard by 2001 or 2002, providing lucrative new royalties. There were also suggestions that TDMA operators in the US such as AT&T and SBC may also dump their networks in favor of CDMA.
The Great Crash Of 2000
Then it all fell apart.
Despite the propaganda and perhaps because of it, the global vendor community had never quite bought into the idea of writing out eight percent of their sales in checks to Qualcomm.
Companies such as Ericsson, Nokia and Alcatel worked on their own version of CDMA, at first given the uninspired name of W-CDMA, but later to be given various monikers such as Universal Mobile Telecom System (UMTS) and 3GSM. NTT DoCoMo, keen to influence and exploit the world mainstream, joined their efforts.
German vendor Siemens, piqued that its preferred variant of CDMA technology had been shunned by its European rivals, went to China and started working with a local telecommunications academy to develop Time Division Synchronous CDMA. Another Chinese start-up, LinkAir, came to Silicon Valley and recruited one of the world’s leading cellular scientists, former Airtouch executive Dr William Lee, to promote its own technology — Large Area Synchronous CDMA.
The object of all these efforts was pretty similar. Take the core properties of CDMA. Then tweak and enhance them so as to put as much technical difference as possible between them and the patents of Qualcomm.
The industry was sending a clear message: “We don’t want to make Qualcomm rich.” And, what’s more, we think we can make a better technology than Qualcomm can. In March 2000, the great tech stock collapse claimed Qualcomm. Its mighty fall, which wiped nearly $100 billion off its value, was so complete that by mid-year some were wondering if a big European vendor such as Nokia might step in and take it over.
The Long Decline
Indeed, as the year rolled on, it became clear that Qualcomm’s prospects were not worthy of that giant market capitalization.
China continued to drag its feet on CDMA deployment. While it gave a smaller operator, Unicom, the right to deploy CDMA networks, at time of writing, there were only four small city nets in operation and one under construction. Tenders had been called again for a national rollout for the second time in two years.
At the Asia Telecom ITU show in Hong Kong in December 2000, China made its true intentions clear.
China’s communications minister Wu Jichuan stated that his country would not pay foreigners for the intellectual property rights used in domestically produced and deployed telecommunications technology. To accentuate his point, uttered in Mandarin, he said that while many people may have different opinions, the only one that counted was his.
Down the hall, Qualcomm CEO Dr. Irwin Jacobs told a closed-circuit TV service that technology transfers of CDMA to China were going to plan and that China would have to pay royalties on its homegrown TD-SCDMA technology.
Indeed, despite Jacobs’ insistence on the patent issue, other vendors continue to maintain that they will have to pay little or no IPR on their own CDMA technologies. For example, many other vendors claimed to have their own IPR for W-CDMA, creating the potential for further legal disputes down the track.
Marginalized
As time went on, Jacobs seemed increasingly marginalized from the world’s industry. In early 2001, AT&T Wireless ignored the pressure to adopt American-style CDMA and elected to go with the European GSM/GPRS standard, and ultimately, W-CDMA. Cingular Wireless — the joint venture of BellSouth and SBC — was heading in the same direction. Again, Qualcomm put on a brave face, issuing a press release that forecast lots of royalties, but by now analysts and shareholders were becoming increasingly skeptical.
Even the CDMA hotbed of South Korea looked shaky for Qualcomm, with its cellular operators confirming that they, too, preferred to go the European/Japanese route for their future network upgrades.
In February 2001, Jacobs ventured into enemy territory — a GSM congress in Cannes, France — and predicted that W-CDMA deployments would be delayed by two years, providing an advantage for Qualcomm’s own technology. Markets were unimpressed, wiping 20% off Qualcomm’s value for Jacobs’ act of talking down the market. Despairing investors would have preferred him to have kept his mouth shut.
It’s not even clear if Qualcomm’s next-generation cdma2000 technology (also called 1X) will beat W-CDMA to market. A claimed 3G network in South Korea operates at speeds of below 154k — and thus, is more analogous to a interim standard such as GPRS than a fully-fledged 3G net.
There were at least two W-CDMA trial deployments — in Britain and Japan — planned for the first half of 2001, and European operators who paid billions for new spectrum have incredible incentives to roll out their networks.
Ericsson insiders have told me that the company is overwhelmed by orders for W-CDMA, which seems to provide it and others with every financial incentive to get the technology shipped.
By contrast, CDMA Development Group chief Perry La Forge told me in October 2000 that the process of balloting members to define the 1X CDMA upgrade would continue well into 2001.
CDMA’s Big Mistake
The big mistake of Qualcomm and its friends was to underestimate the power of GSM and its ability to give operators what they wanted.
The selling point of CDMA is its capacity levels, which allow operators to support more users on existing spectrum and to more easily upgrade to higher data speeds, if and when demand for this eventuates.
But the selling point of GSM is its rich selection of standardized features that provide real incremental revenues. Capabilities such as SIM cards, short messaging and roaming increase customer value and operator revenues. When standardized across countries, GSM also benefits from tremendous economies of scale (it outsells CDMA by a factor of four to one). This means that GSM handsets continue to benefit from increased range and superior form factors.
Qualcomm’s view of the market was colored by its American origins. Like its TDMA counterpart, it rather lazily grafted its CDMA interface on to the old analog signaling standard. It also downplayed the need for solid, standardized feature sets.
When Qualcomm and other CDMA vendors finally realized the advantages of GSM’s features, their reaction was too little, too late.
Qualcomm’s prospective CDMA customer in China, Unicom, had been stating publicly for two years that it wouldn’t adopt CDMA until the platform supported GSM SIM cards that would allow it to churn across existing GSM customers with ease. But it wasn’t until late 2000 that such a solution was announced — at the joint initiative of Unicom and smart card specialist Gemplus.
A similar initiative to create roaming was finally developed by Asian operators in 1999, seven years after it became commonplace in GSM. This was a particularly grave failure by the CDMA industry. GSM networks generated as many as 8 billion international roaming calls in the year 2000.
In markets where CDMA competes against GSM, CDMA often comes off second best. India’s MTNL, after a careful evaluation of CDMA’s lack of features, was moved to brand its planned service as the “poor man’s phone.” In Singapore, the technology was offered as a discount service and then simply withdrawn when regulators requested that the network shift frequencies.
Qualcomm’s capacity for hype is one thing yet to change.
In early 2001, Qualcomm forecast it would sell 90 million chipsets for the year. After seven years of commercial operation, CDMA had less than 90 million customers in total worldwide. CDMA adoption rates would have to accelerate to record levels for Qualcomm to meet those estimates.
By contrast, GSM had reached 452 million people globally by early 2001 — about one handset for every 12 people on earth. Some 200 billion data messages are forecast to be sent to and from GSM handsets in 2001 — 40 for every human on Earth. By the middle of 2000, GSM users were generating some 20 million international roaming calls a day.
Between Sep. 2000 and Jan. 2001, GSM added more users than CDMA had gained in its first six-and-a-half years. GSM’s victory was absolute.
Grahame Lynch is the former group editorial director of America’s Network and is now CEO of Decisive Media in Australia and SE Asia. <<
... and one more:
>> GSM Primed for US Takeoff
Grahame Lynch
Americas Network
March 2001
Amid the flurry of product and strategic announcements this week at the CTIA Wireless show in Las Vegas, one may prove particularly significant for the future dymanic of the US cellular industry. Siemens' planned tri-band TDMA/GSM phone is just exactly what AT&T and Cingular Wireless need if they are to successfully migrate their customers from their TDMA networks to newer GSM networks with packet overlays. And the greater economies-of-scale inherent in GSM handsets should enable AT&T, Cingular and VoiceStream (the major North American GSM provider) to compete with the best of CDMA's cutting-edge offerings.
TDMA and CDMA are running neck-and-neck in the North American cellular race with 31 million and 29 million customers respectively. GSM, which is currently locked into the non-conformist 1900 MHz band in the US, has just 9 million customers in North America (most GSM nets operate at 870-900 MHz).
Globally, GSM's 441 million customers swamp TDMA's 61 million customers and CDMA's 80 million customers. Reports coming out of CTIA this week suggest that both AT&T and Cingular may move more quickly to GSM than expected - AT&T originally said it would push GSM's packet-based "GPRS" overlay in its 1900 MHz network only, but some speculate that it may also overlay the technology on its major 850 MHz band network.
The key attraction of GSM isn't the air interface as such but its underlying network system which provides a rich stream of standardized features such as international roaming and short messaging. Although CDMA nets support these features they tend to be more problematic due to a lack of global specification - for example, the CDMA standard doesn't provide for a numbering plan that can support seamless, automatic global roaming.
In a statement of intent as to where it sees its next market beach head, the global GSM community will descend on the United States to TDMA's heartland - Seattle - later this month for its 2001 plenary. <<
Best,
- Eric - |
