The Bleedin' Edge:
EDGE - a definite for North America, a faint maybe for elsewhere
27-Jun-2001
By Julian Herbert, Managing Director
At IBC's conference on 'EDGE in a 3G World' in Dublin, 5-7 June 2001, a relatively thin but enthusiastically probing audience sought answers to some fundamental questions: will EDGE be a technological reality? Have operators made incontrovertible commitments to EDGE networks? When will infrastructure be rolled out and most crucial of all, will handsets appear? Presentations were made by a number of major GSM infrastructure manufacturers including Alcatel, Lucent, Ericsson and Nokia, several operators, industry associations and industry watchers. There was broad consensus that EDGE could and would be realised technologically.
There were firm commitments to roll-out expressed by two North American operators, AT&T Wireless and Cingular Wireless, but there was ambivalence and even silence from key manufacturers on timescales for handsets. At the end of two days of frank discussion, the audience was still left pondering an age old question: which comes first, the chicken or the egg? Handset manufacturers said
that they would mass produce once operators had committed, and operators said that they would commit when they could see working product. At least, it was concluded, the talk centred on whether EDGE would be a reality, and not whether EDGE was dead and buried. [Cold comfort]
John Hoffman, who spoke on behalf of the GSM Association and
chaired the first day, was clear about the GSMA's intention to facilitate the adoption of EDGE by its members, but was equally clear that the GSMA facilitates and promotes to, but cannot speak for the intentions of, its member operators. His concluding statement summed up the GSMA position: 'It is time for EDGE leaders to stand up and lead. Tell operators, tell manufacturers, tell the financial community [about EDGE].' He also called for the
honest management of customer expectations around both GPRS
and, if adopted, EDGE. This point was returned to repeatedly. The fact that data rates range from 30Kbps to 384Kbps, (depending on propagation characteristics of the deployment band, distance from the centr e of the cell, whether Classic or Compact is used and coding schemes), indicate that this issue is, frankly, best left out of
any consumer marketing literature. [That's right, snooker the suckers. Again.] There are too many variables
and too much uncertainty.
Airtel was one of a handful of operators represented on the
programme and its position was that of interested evaluation rather than outspoken advocacy. Airtel concluded that a coverage scenario where EDGE was deployed in parallel to UMTS would be unprofitable. Its deployment rests on: the enhancement which the
GERAN (GPRS/EDGE Radio Access Network) would provide to
GPRS if the timescales for UMTS are pushed further back and the possibility of EDGE offering complementary coverage to UMTS in rural areas. Perhaps the most important statement was the argument that whilst in ideal circumstances EDGE offers better spectral efficiency than GPRS, there are no bit rate differences over 100% of a GSM cell. Further, it was argued, because of limited C/I improvements and the fact that frequency hopping is not allowed when EDGE transceivers are deployed in GSM/GPRS frequencies, EDGE is better deployed on dedicated frequencies, as an optimised data service. [Spectral efficiency? Well...] Whilst this may be the best way to enhance data rates, it does not give the necessary spectral efficiency gains.
A paper from Lucent illustrated lukewarm support. The
representative spoke of a 'lack of conviction' amongst people in his own team, let alone in the wireless industry at large. [You can fool some of the people all of the time..] More base stations are needed to provide equivalent coverage to GPRS whilst maintaining data rate or capacity gains, the business case hinges on the timing of UMTS and the case for end users wanting to pay a premium for higher speed service is unproven. The rather gloomy conclusion in this case, was that EDGE may have a role as a proving ground for user acceptance of higher speed services and as
a USA-specific in-band 3G technology (for those wanting to migrate through GSM/GPRS), but that the development of optimised applications is crucial. Further, EDGE will not be universally accepted by the GSM operator community. [What's the choice, pay for the spectrum and cross fingers wrt GPRS?]
GAIT
The UWCC restated their unequivocal support for the convergence of TDMA and GSM via EDGE. This raised separate questions about GAIT, strategy for the 850MHz band and whether EDGE had a future outside the constituency served by TDMA operators in North and Latin America. The two US operators represented, AT&T Wireless and Cingular Wireless, were unambiguous about when they wanted EDGE to be a reality for them - as soon as possible. The emphasis is shifting away from EDGE as a 'convergence enabler', towards the use of EDGE to improve and enhance GPRS. Indeed by the end of the two days, the nomenclature of currency tended towards EGPRS and not EDGE.
Further strength to this argument was provided by Cingular
Wireless. It appears that Cingular will overlay GSM in the 850MHz band, so the long term necessity for Cingular to use GAIT is doubted, since the network may ultimately be 100% GSM. AT&T Wireless on the other hand, is still, officially at any rate, behind GAIT and wants handsets in the marketplace by the end of 2001 or early 2002. [AWE was clearly the dunce at the party.] Once again an absence of comment on timescales from handset manufacturers who were present failed to make it clear that GAIT will have a role to play. The impression was that the best that can be expected is the delivery at the end of 2001 of a 'two phones in one' cludge from Siemens, with a model with genuine interworking by mid-2002. Ericsson and Motorola are also thought to have a working product.
Necessity for GPRS to be successful
If EDGE is to have any chance at success, GPRS must prove itself as a viable first step. Whilst there was acceptance by most speakers that GPRS was a 'done deal' there do still exist questions over the volume of handsets and users which can be expected in the market by the end of 2001. GPRS is essential as a proving ground because it will demonstrate a number of key factors: that operators can manage complex packet data overlays, that operators have available and can market services which use them, that operators have resolved billing issues, that customer expectation has not been misguided and that the handsets (which represent less complexity than EDGE handsets), are stable and can be produced in volume and sold at prices which are affordable to both operators and consumers alike. [Nokia better deliver this Fall or it's match point to Q.]
There was no disagreement on the issue of speed: i-MODE has
demonstrated that there is money to be made out of very low speed data, provided applications are optimised and customer expectation is managed. i-MODE delivers ARPU of around $20 a month with speeds of 9.6Kbps and the Blackberry RIM service in the USA delivers a premium service with speeds as low as 7Kbps, because of efficient optimisation. On this point it was grudgingly conceded
that the GSM community has lost ground in the PR war to the
CDMA lobby, which is already demonstrating CDMA 1x networks
with data rates of up to 144Kbps (optimal) running over
commercially available handsets. [Not PR. Reality] The message must now be clear: promote what GPRS can do (provide services which users want) and not what proponents were once claiming it could do. [Ooops,so we over-promised a bit. Sorry.]
A paper from Northstream stressed this point most emphatically: EDGE will not realistically achieve data rates of 384Kbps and GPRS will not achieve rates of 115Kbps. [Uh, oh. It might be match point to Q already.] In fact using coding scheme 6 and four time slots (deemed realistic), EDGE could achieve data
rates of between 100Kbps and 120Kbps, which would degrade
according to load and interference (or distance from the cell-centre). [That's a lot of expense to straped carriers]
The major driver for EDGE in these circumstances is as a network which provides broad coverage for 3G services. The sticking point remains, however: no handsets, no business case. [Oh, my. Tut, tut]
Cingular Wireless and AT&T Wireless
Both these operators were unambiguous about their ambitions for deploying EDGE. Both want to deploy and both want to deploy during 2002. Cingular's legacy of GSM-1900 in former PacBell and BellSouth DCS Mobility properties in California, Nevada and the Carolinas, and TDMA in its BellSouth and SBC cellular properties, has meant a legacy for which EDGE was ideally suited. GPRS is already being trialled and it will have a strong focus in properties on
the east and west coast. Factors determining Cingular's choice of EDGE were cited as: GPRS throughput limitations, concerns over 3G spectrum licensing [Oh, that is choice. Thank you, D. C. Circuit Court of Appeals. Let AWE stew.] a need to provide a competitive data rate, technology legacy, existing vendor support and cost. The limitations
of GPRS are based on the continuing growth in voice traffic which the operator is experiencing, which will likely restrict the amount of spectrum dedicated to data traffic to a maximum of two time slots; this, coupled with the fact that GPRS equipment will, in general, only be supplied with support for CS-1 and CS-2, means that the theoretical maximum data rate for GPRS will be 52Kbps. EDGE will
provide this with a single time slot. Interestingly, the Cingular representative was outspoken about commitment which the operator had received from vendors for trials of EDGE-enabled terminals by Q3 2001. [Nokia hype?] While there remains only one vendor (Motorola) shipping a stable GPRS handset in volume, there was a less than tacit
acceptance that this date would not be met. [No, Nokia is being taken with a grain of salt?] A final comment hammered another nail in the coffin of UWCC-136: 'Nothing I have talked about includes TDMA,' suggesting that Cingular's strategy for the 800MHz band may be based on the roll-out of a GSM-800 overlay. There was no firm date given for EDGE availability. This depended, according to the speaker, on vendors and handsets.
AT&T's presentation was a good deal more zealous.[Read: Almost demented.] 'A few years out, like it or not, you [operators] will not be able to buy GPRS-only radios. They will all be EDGE capable,' declared the speaker. This summed up the message: EDGE is simply an inevitable part of the development of the GSM standard. This message was delivered with a firm roll-out plan and a clear statement that multi-mode multi-band GAIT forms a key part of the migration strategy.
The role of MVNOs
One part of the potential business case for EDGE, for operators with sufficient spectrum, is the allocation of frequencies to MVNOs. Microcell Connexions, the service provider arm of Microcell, the GSM-1900 operator in Canada, provided a reasoned argument for pushing penetration beyond 100%, using market segmentation with MVNOs. 'The last thing a wireless operator needs today is another capital investment opportunity,' came the opening statement. [I love this guy; he's nicely ironic] With ARPU declining and the only opportunity for an operator to acquire customers being churn, the industry is facing 'a market moment'.
EDGE (or any other data-enabling technology) provides the
opportunity for the use of spectrum by data-dedicated services provided by MVNOs.
There are four types of MVNO, which could dedicate arketing and sales effort to specific data services: operators themselves, alternative operators (fixed, ISPs, ASPs such as Energis, other companies such as AOL/Time Warner), horizontal consumer brands (Virgin), professional user groups (public utilities, public services such as emergency services). Two examples were cited of successful MVNOs in the USA. OnStar, a value added services provider for General Motors, which buys airtime from Verizon, is sold
as a package to purchasers of new GM cars. The gross margin on the sale of the car alone over four years is estimated at around USD 500; the additional gross margin provided by the OnStar service package is USD 1,500. The other example is the Blackberry service, which costs USD 60 per month and USD 800 for the device, runs at just 6/7Kbps, but provides a very adequate user experience.
The argument which applies to MVNOs running dedicated EDGE
networks can of course, apply equally to GPRS or any other mobile data technology, such as CDPD, Mobitex or CDMA 1x.
Handsets - the chicken or the egg?
The conference room remained full for the final session. TTPCom provided a simple illustration of the challenge which lies ahead with EDGE terminals: using imagery from the circus, the audience was asked to consider that developing GPRS terminals was analogous to learning to juggle a very large number of items - it is highly complex; the move to EDGE was analogous to the same person learning to lift a heavy weight, a task of much less complexity, but
far more difficult. The reasoning for this is that for handset manufacturers, the additional challenges for EDGE, once GPRS has been mastered, are focused on the modem components at DSP and RF level, because of the increased complexity of equalising the RF in EDGE. Techniques to overcome this (using feedback and/or pre-distortion) increase the processing requirement in four time-slot
EDGE to 400 MIPS (from about 100 MIPS in GPRS). Despite this, TTPCom argued that as silicon componentry represents just 15% of the Bill of Materials (BOM) cost, even a 50% increase in silicon would only result in a 7.5% increase in BOM; it is the contribution of 'low tech' components, such as colour displays and better batteries, things vital for new applications, which will lift the BOM by an estimated 10% to 20% over GSM/GPRS. EDGE handsets will offer therefore, three times the data rate of GPRS with equivalent
time-slots, identical quality of service and just a marginal increase in product cost.
As an indicator of the opportunity which EDGE has: UMTS, by
contrast, will require 5,000 MIPS. It will initially offer three times the data rate of EDGE, controlled quality of service (ie, better in some circumstances), but mean significant increases in product cost. 'It will be five years,' maintained the speaker, 'before we can achieve
cost parity between 3GPP [UMTS] and GPRS/EDGE handsets'. As a sideline comment, there was another call for managed user expectations: 'WAP was indeed the killer application; it was the application which killed consumer demand for data services.' The following speaker, from Philips components, reiterated this message. EDGE will be available in 2002/2003, provided the semiconductor industry can reduce cost, size and power consumption and increase features.
Handset and semiconductor representatives then joined a panel to discuss whether EDGE will be a reality and if so, when and where it will be deployed. As with the previous two days of discussion, there was plenty of warm sentiment and positive noise, but no firm expressions of commitment. [Lots of BS, in other words.]
During the discussion, Nokia expressed confidence that there will demand for EDGE service, first in the USA and Canada, but called for any expression of interest, let alone commitment, from a GSM-900/1800 operator. [Read: Nokia's not pushing it until the carriers commit.] Demand will be driven by MMS from 2002 and by the possibilities offered by Java-enabled handsets. The message again, is that speed (data rate) is not of the essence. [We love a slow cludge.] Ericsson said that it was in discussion with a number of operators over terminal delivery, expressed a 'clear commitment' to EDGE and said that there were encouraging signs of operator participation [They're finally returning our phone calls.] [, especially in Europe. Nokia and Ericsson both insisted that they will deliver working EDGE terminals, but no timescales were given, despite very direct and frustrated questioning on the issue by Cingular Wireless. Tropian presented the conviction that EDGE will happen in the USA and Canada, but that in Europe it will depend on user take-up and the speed of user take-up of GPRS. EDGE is technically highly achievable and it has a future, provided the cost delta over GPRS is very minimal. Both TTPCom and Tropian insisted that the way to heighten operator interest, currently nonexistent outside North America, is to build and demonstrate hardware which can in turn demonstrate money-making propositions, which are not possible without EDGE.
Handsets in their first incarnations are likely to support asymmetric data transmission, with 2+1 or 3+1 timeslots.
SEE LINK FOR ADDITIONAL CHARTS AND INFO
e-searchwireless.com. |
