re: EMC on Historical Parallels between GSM & WCDMA Evolution
A Quacking Quidiot from the Qualcomm Hear No Evil, Speak No Evil, See no Evil thread wondered why this wasn't posted here.
In this article Julian Herbert, EMC's Managing Director, examines historical trends in the wireless industry, and draws parallels between GSM & WCDMA.
There are 3 charts embedded (.gif files) in this article. The first is particularly interesting. Right click on the link for each to save to disk.
A rather complementary article dealing with the same subject matter was published in the Financial Times on the same day that Julian published this article:
Message 16706851
>> History Shows GPRS/W-CDMA Delays 'Absolutely Predictable'
Julian Herbert
EMC Cellular
November 21, 2001
e-searchwireless.com
Whilst it is by now obvious and a little passé to point out that new technology is never delivered on schedule, and to draw comparison between GSM in the early 1990s and GPRS or W-CDMA today, the history lesson may give clues as to how late the delivery dates may eventually be.
The following quotations, expressing concern over delayed implementation of new technology, are indistinguishable in tone, but separated in time by more than 13 years. So similar are they, that when asked to call out a date, an audience at the UMTS Congress in Barcelona (October 2001) dated the second of them no earlier than 1997.
'A stable standard...is not expected till the middle of next year. European operators have to wait while it evolves. New components in handsets and the complexity of ... networks ... represent headaches for manufacturers and engineers.' - Chris Nuttall, 18 September 2001 in FT Marketwatch summary of CSFB report on '10 reasons why 3G is delayed', -
'Incomplete recommendations, a tight timetable, new advanced mobile radio techniques embedded in a complex system represent both risks and a challenge for ..[our]...industry'.
Analogue-Digital Migration, 1991-1996
Looking at the world subscriber growth trend, both analogue and digital, plotted since 1983, the analogue curve shows almost perfect symmetry in its rate of increase and decrease. Digital shows very low rates of growth over the first four years of its history. The figure below highlights four points of inflection during the early migration process.
[Chart of Migration Trends of World Cellular Subscribers from 1983 to 2000 plots the four inflection points below]
e-searchwireless.com!OpenElement&FieldElemFormat=gif
1) Between 1990 and 1992, in the months leading up to the introduction of GSM in Western Europe (which saw the introduction of competition in Western Europe), incumbent PTT operators invest in the improvement of their analogue networks. They are either configured for capacity or for handportable use. The superior coverage which analogue enjoys over digital, gives analogue growth a boost. GSM is launched in mid-1992, in France and Germany, a year later than the MoU signatories had pledged, and makes little impact on subscriber numbers for a year.
2) By 1994, GSM has launched in most Western European markets, bringing with it claims of superior quality, international roaming and encryption. The coverage advantage, combined with the increased advertising and general profile which cellular competition has been given, means that analogue subscriber bases are growing faster than ever before. Digital is making an impact in markets whose analogue technologies have saturated (France, Germany), but in NMT-900 and E-TACS markets (UK, Nordics, Italy), GSM is still flagging.
3) By the end of 1996, four years after the launch of the first digital networks in Europe, digital finally overtakes analogue in terms of net additions, largely through a proliferation of network launches in Central and Eastern Europe, the launch of CDMA in the USA and Far East, the accelerated take-up of PDC in Japan, and the relaunch of TDMA as IS-136. Analogue still does not decline and grows on a world basis as much in 1996 as it had in 1994.
4) 1997 sees analogue reaching the top of the S-curve and digital technologies overtake analogue in terms of both net additions and cumulatively. It has taken over five years for digital technologies to reach a global consumer market and to displace prior generation technologies.
The 'GSM Parallel'
In summary, it took standards' bodies, manufacturers and operators, six years to realise the firm 'Pan-European Digital Cellular' concept first embodied in the creation of the permanent CEPT nucleus (Groupe Spéciale Mobile or GSM) in August 1986. It took a further five years after launch, for new generation digital technologies to displace analogue globally. A large part of the slow pace of change must lie in the totally unforeseen positive effect which the launch of competing digital systems was to have on analogue subscriber growth. This gives credence to the idea that it is not necessarily technology which sells, but how assets are positioned, marketed and sold.
A simple parallel between GSM in its early days and GPRS indicates that there is nothing surprising about the absence of a mass market at the moment: GPRS was conceived in 1994 and standardisation work completed, five years later, in 1999 (with GSM release 99). The standardisation work enhanced an existing standard, rather than defining it from scratch, so five years rather than the six which GSM took, was predictable. A mass market for GPRS, using the GSM parallel, will not be achieved until 2003 at the earliest.
The first vision of what UMTS might become saw the light of day in the form of a RACE Workshop Report in early 1995, but standardisation work, it could be argued, did not begin until after the UMTS Task Force presented its results in mid 1996 and ETSI's subsequent UTRA decision in January 1998. The GSM parallel would see UMTS (W-CDMA) realised at the earliest by mid-2002 and possibly as late as 2004. Given the greater complexity of UMTS over GSM, slippage beyond 2004 is both predictable and, in historical terms if not investment terms, quite acceptable. Indeed if the parallel begins at the point when there is a stable group dedicated to standards' development, then the start must be the inaugural meeting of 3GPP in December 1998. Six years from this date would put commercial reality at early 2005.
Second Generation Digital: A House Built On Rock
One of the clearest early articulations of what GSM was to become was presented by Thomas Haug at the second DMR conference in Stockholm, October 1986. He stated that the two major advantages of a Pan-European network would be international roaming and a Europe-wide market. Principally therefore, advantages were perceived as improving the market for basic mobile voice telephony, a market which had been proving itself in the Nordics for nearly five years at that date.
The first working party schedule published in July 1987 by the CEPT GSM, bore this out, specifying the need for work on data transmission at 'standard PSTN rates' of 9.6Kbps (circuit switched) and for 'bearer services' most of which were concerned with enhancing the experience of voice telephony (voice mail, call waiting and other services). Packet data was to be used for signalling purposes only, part of which was to be used for low priority point to point and cell broadcast messaging, in other words, SMS.
By the time GSM launched in Europe in mid-1992, that there was a market for simple voice services was proven. The industrialised world had been using public switched fixed networks for over 100 years and mobile enhancements to it, for over a decade. Systems which could make the concept more genuinely mobile (handportables) or more accessible to a wider market (lower prices through competition, economies of scale) were bound to be successful.
It is with the 'mould breakers' which arrived alongside GSM in Europe, that the concept of mobile voice telephony entered the mass market at a rate and in volumes hitherto unprecedented. Changes to competitive environments, the end of PTTs and telecom monopolies, tariff regulation, the ways in which companies attracted and interacted with customers and the marketing of handsets, was quite unlike anything which had happened before, but was still predicated on a simple service offering: voice telephony combined with mobility. Data, messaging and bearer services were to be enhancements, but not central planks of the standard, so whilst the late nineties boom in SMS has been welcomed, it was certainly not expected or planned.
2.5G/3G: Houses Built On Sand?
The business case for later generation services, GPRS and W-CDMA, assumes that there will be demand for data. Early discussions in the GSM community about the need for a higher rate data service were prompted, not by predictions of mass market take-up, but by evidence of a growing number of dedicated professional mobile data services, such as Mobitex, across Europe. Services in mind, according to SMG worksheets in 1994, were 'fleet management, logistics, telematics and mobile offices'. Hardly applications with mass market appeal.
The availability and success of mass market fixed internet applications and the 'dotcom' boom, spurred some handset manufacturers from about 1999, to claim that the internet was now mobile. Whilst IP interworking had been specified in IS-95 (CDMA) since the early 1990s, it was not completed in GSM specifications until release 2000 (February 2000). Leaving aside the fact that WAP was a marketing disaster, not because it did not work, but because its proponents made unrealistic claims for it, the experience illustrated one thing: that bringing the benefits of IP to mobile would mean tailoring it to the specific needs, and limitations, of the mobile user. The business case for 2.5G and 3G networks has always been based on the assumption that demand for data simply exists. Absence of firm evidence has been repeatedly observed over the last several years by EMC:
'Questions concerning the actual market demand for mobile data services went largely unanswered.' (EMC, reporting conclusions of 1st UMTS Conference, London October 1997).
'There is clearly some level of uncertainty about how the general public will embrace the varied service applications that will be available with the roll-out of UMTS networks.' (EMC, reporting conclusions of UMTS Congress, Barcelona October 2000).
Where data services have been successful, they have been limited to quite specific markets, notably i-Mode, BlackBerry and Mobitex. If you add to this the burden of debt created by licence auctions in the UK and Germany, the level of expectation, placed on non-voice services, is unacceptably high. Whilst ARPU from non-voice services is probably increasing, it is taking an increasing share of at best stable, and at worst declining, overall ARPU.
Crumbs Of Comfort
There are some crumbs of comfort in all this gloom. Uptake of mobile data services is high in certain markets, where applications have been optimised for handsets specifically designed to enhance the user experience of sending, retrieving and exchanging information. This amounts to simple services delivered on handsets with colour graphics displays and acceptable power consumption levels.
e-searchwireless.com!OpenElement&FieldElemFormat=gif
Network operators in Japan and Korea are launching services which are aimed at the mass market, including messaging services such as logo and ringtone download, photo-messaging, games, chat and email. Whilst it is too early to tell whether this will have a positive effect on ARPU, there is a proliferation of Java-enabled handsets featuring large full colour graphic displays and built-in cameras, such as the Samsung SCH-V200 for cdma2000 and the Sharp J-SH07 for PDC. Whilst cdma2000 and PDC networks are leading the way, the applications which work are technology independent; if they find consumer acceptance in these markets, there is no reason why some of the applications, enabled by the availability of the right type of device, should not find an even wider market in the GSM/GPRS/W-CDMA world. Given the relative size of the addressable market for GSM, against that for CDMA, there is no reason why economies of scale should not be achievable. History is demonstrating, however, that arriving at that point, may yet take a further five years.
[Chart GSM/WCDMA v. cdmaOne/cdma2000: Current Addressable Market as of June 2001]
e-searchwireless.com!OpenElement&FieldElemFormat=gif
GSM Markets with UMTS licenses 363 million
Other GSM Licenses 183 million
Total GSM 546 million
Markets with cdma2000 Deployed 81 million
Markets with cdma2000 Deployed 16 million
Total CDMA 97 million
###
- Eric - |
