To the Future and Back
The issue of inter-generational roaming is fraught with unanswered questions and unresolved technical challenges, says Elizabeth Biddlecombe.
International roaming has contributed enormously to the success of GSM, delivering service to users across continents and boosting mobile operator revenues. Yet only now, with third generation mobile services imminent, is roaming between second-generation standards beginning to be enabled. The real challenge for roaming - that of crossing the boundary between present and future mobile generations - is still to come.
In contrast to GSM, where roaming was introduced some years after the technology itself, third generation systems have held global roaming and backwards compatibility as a central tenet from the outset. In fact, there are a number of reasons why 2G to 3G roaming is a commercial imperative, as much within country borders as without.
For a start, GSM will be around for some time to come. Ericsson expects that it will be in use until 2008 or 2009, while Siemens goes further, suggesting GSM will survive until 2010. Initially, third generation networks are going to be deployed in urban, hot spot areas. As the user leaves those areas their mobile communications will run over an existing 2 or 2.5G network.
'Inter-generational' roaming will then take place either within one operator's network, or - more controversially - competitively between the incumbent operator and the new entrant to the market. Needless to say, national regulators will be keeping a watchful eye to ensure a level playing field. For instance, both the UK's Oftel and the CMT (Comision del Mercado de las Telecomunicaciones) in Spain, have inserted a condition into current licences mandating that the incumbents must allow a newcomer to roam onto their network. Both regulators have the power to intervene if agreement cannot be reached.
COMPULSORY ROAMING?
Some operators are critical of obligations of this kind. Neil Montefiore, CEO of Singapore mobile operator M1, feels that such regulation will be acceptable if there is a reasonable return for the existing operator, but not if it is regulated at a low rate. "Initially, I think it will allow new entrants to concentrate on providing 2G services to 2G handsets, rather than building out new infrastructure. 3G customers will focus on the 2G business because that is where they can get the customers. If the regulator's intent is to introduce new technology, then it is counterproductive to insist on roaming."
However, Oftel for one has included measures to ensure that TIW, the new entrant in the UK market, is making adequate progress towards building out its own network before it can roam onto that of its competitors.
Then there is the more unlikely scenario whereby a 2G operator might negotiate a roaming agreement with a 3G operator, perhaps to make use of spare capacity on the 3G network or to resell 3G services as a service provider. In general, however, there is not much of a commercial imperative to drive the market towards roaming if you are a 2G operator, argues Alan Pyne at London-based consultancy Schema - unfortunate in light of his observation that "it will be technically complex and require a higher degree of cooperation between network operators than has happened seen so far."
ADVANCE FORGES AHEAD
But it won't always be two competing operators tackling these technical complexities. Alliances such as the BT and AT&T wireless joint venture, Advance, will also have a need for 2G to 3G roaming. In July, Advance unveiled its first products to allow business users at least basic service on the same phone number while travelling between Europe and the US - the WorldConnect roaming service and the WorldView single bill service for phones being used on both sides of the Atlantic.
Advance's WorldConnect product will initially bridge the TDMA (IS-136) and GSM gap via plastic roaming (a solution that lets the user's SIM card be swapped to a new phone that works on the correct air interface), with multimode handsets expected before the end of Q2 next year. Andrew Noble, head of mobile multimarket products at BT, explains that implementing these services was challenging. "Signalling and commercial issues were difficult, as was billing. We needed something that translates the records quickly. You can't afford for it to take a week to get a record from one side of the Atlantic to the other. Fraud levels are very high in the international traveller sector," he said.
BT and AT&T Wireless Services are already looking at converging on an IP core to enable a smooth transition to GPRS and EDGE. And the two are engaged in a GPRS roaming trial, along with Taiwanese operator FarEastone and SmartTone (in Hong Kong), though Noble says that as yet there are no conclusions to draw as far as service translation is concerned. WorldConnect enables only voice services to travel at present, and while Noble acknowledges the importance of service translation, he reminds us that, "voice is still a killer application, and will be for years to come."
The Global Roaming Forum is also focusing on GSM-ANSI 41 roaming as part of its mission to map GSM to other 2G technologies around the world (including iDEN and TETRA), though it started too late to be of assistance to BT and AT&T. The Forum is led by the GSM Association, but includes the CDMA and TDMA communities (represented by the CDMA Development Group and the Universal Wireless Communications Consortium). This work has been driven by the North America GSM Association, reflecting the necessity for roaming between TDMA and GSM in the Americas. Meanwhile, in Dublin, Logica Mobile Networks has already developed ANSI-41 to GSM MAP gateways.
While the subject of 2G to 2G roaming seems a little dull when 3G is in the offing, the Forum is laying the groundwork for 2G/3G roaming, and has input from both NTT DoCoMo and the UMTS Forum, as representatives of the 3G community. However as Eric Hill, director of inter-standard roaming at the GRF, says, "we are concentrating on 2 and 2.5G, being mindful that everything we do has to feed into where 3G is going. But we are walking before we run."
A WORLD OF DIFFERENCES
But technology interoperability is not all there is to roaming, as Eric Hill of the Global Roaming Forum explains. "We are not only developing technologies, but also principles for commercial arrangements so that there is a graceful migration path."
"We are going to agree a standard roaming template over the next six months, which will lay the groundwork for 2G/3G agreements," says Eric Hill. Such guidelines will also have relevance to intra-3G roaming, as there will be operators running W-CDMA on a GSM core who need to interconnect with others running W-CDMA over ANSI-41.
For the first time, roaming negotiations will have to make mention of the welter of services expected from 3G - what Nokia calls the 'hypermarket' of applications. "The question is: what is not a basic service?" says Eric Hill. "What constitutes more than the bare minimum expected by the customer?"
In practice, what constitutes a basic data service is defined by technology constraints. But even when a set of common data functions has been defined, there is the issue of how to share the revenues, as Cluster Consulting's Ben Chesser points out: "Discussions centre on how to split the money (from content). There is no consensus regarding billing per minute or per megabit."
Neil Montefiore at M1 is worried by the lack of cooperation in this area, and has proposed the topic for discussion at the next plenary session of the GSM Association in October.
"There is not enough work going on," he says. "The next two years is not a long time to get hundreds of operators to agree. It is the customer experience that will determine the success of the technology - and customers will expect that the services will follow them around the world.
"Otherwise, they will find what we are offering inferior to 2G voice."
TRANSLATION DIFFICULTIES
The translation of services brings a new layer of complexity to the question of 2G/3G roaming. Services will have to be reformatted to cover all the variants available in the target market - for all the different flavours of 3G, including not only differences in standards, but the differing approaches to packet loss, for example, between Ericsson and Nokia.
Varying 2.5G implementations must also be accounted for as well as specific handset capabilities. "Ultimately, you have to be prepared for GSM, GPRS, UMTS, Wireless LANs," says Kari Penttila, VP of engineering at independent value-added service provider Iobox. "It is complex," he adds, "though at present we don't know yet what services will need high bandwidth. Nobody can tell, so we designed our platform so we can act quickly."
Services will not only have to be adapted for varying transmission technologies.
Surrey, UK-based Argo Interactive develops transcoding software for converting information rendered in the web's HTML into wireless mark-up languages.
Transcoding between wireless mark-up languages - HDML, C-HTML and, of course, from Nokia WML to Ericsson WML, for example - will be the next step.
"The browser may only be compatible with a mark-up language not supported by the visited (2G) network," says Richard Jelbert, CTO. However he points out that even with transcoding there are more problems. "The device may be able to read the page, but you may not be able to read the language. Natural language conversion is another issue," he says.
But whether information services will be supplied by the visiting network isn't clear, as many operators will want to keep their branding in view and route traffic through their home page. As Stewart Scott, business development manager at mobile services developer Airflash, explains: "It is hard enough to get users to adopt information services on their mobile.
The people who establish a brand will gain trust. Users are not going to churn from that provider in a hurry and will want to take that brand abroad. The ideal situation is that the user gains access to services in the home country from the network bearer."
Scott points out that operators derive the majority of their roaming revenues from a handful of companies, so that it would be feasible for the home operator to develop information specific to those countries, with a knowledge of the parameters at play on the visited network.
However, Neil Montefiore at M1 thinks that the global mobile portal isn't the way to go forward. "Vodafone, MSN, Yahoo are all trying to take the internet portal model and transfer it to the wireless world. But usually it is a local portal that gets the most visitors on the internet. Operators should agree on a suite of content services that the customer selects in their home market, but can access via an identical process abroad. Operators could share information or provide links into the home network portal. There are lots of ways of doing it. We need to start trying."
Apart from content provision there is the issue of service management in general, as Jussi Koski, development manager at Sonera, points out: "How is the customer told that they have roamed onto another network and therefore the capabilities have changed? How would we alert them that the price of a particular service might have changed now they are on a 2G network?" he asks. He believes such questions are hard to resolve when the performance of GPRS on a fully subscribed live network is not yet known.
TERMINAL ADAPTABILITY
But all this talk of services is academic without handsets. Vendors claim that dual-mode, 2G/3G handsets will be available from the outset in Europe (around 2002), but in Japan, single-mode handsets will only be available where 3G services launch next year. Handsets will incorporate at least one 2G and one 3G technology, though in the long term more than one 3G technology (for instance W-CDMA and UTRA TDD) will be deployed by an operator, in order to make the most of spectrum awarded in the licence.
As a spokesperson from one manufacturer pointed out, this creates an engineering challenge for handset designers, given the current trend for smaller and smaller phones. "There are so many technologies to incorporate that, until integrated chips arrive - which may take five or six years - phones will have to be bigger," he said.
Ahti Vaisanen, VP for Nokia's W-CDMA business programme, says that, "adding GSM to W-CDMA terminals is not going to be easy, but it will be a simple task compared to W-CDMA terminal development - a big challenge in itself." As for the increased size of 3G terminals, he isn't worried, pointing out that there will be "a lot of new concepts, products such as imaging phones, communicators etc, where the display size is of great importance.
When concepts change it is difficult to compare sizes. But given reasonable time - two-to-three years - the size difference between GSM and W-CDMA functionality will be negligible."
Apart from the reappearance of the brick form factor, there are more basic concerns. As yet it is not clear how the handover between the 3G and 2G network will happen. Ahti Vaisanen summarises Nokia's approach: "The handset measures signal strengths of both networks continuously, communicates to the system that it is losing the W-CDMA signal and needs a GSM channel, for example, and the system allocates a GSM channel." He does not believe that users will need to swap phones, though a spokesperson from another handset manufacturer said that this was likely in the early stages. Whether those early adopters who venture into 3G will stomach this is another matter, adding to concerns that users will not tolerate the need to change handset every six months in order to take advantage of mobile data developments. Indeed, Ben Chesser at Cluster Consulting suggests that, "it is going to be hard to sell 3G handsets when you only have 40 per cent coverage".
The issue of 2G to 3G roaming is clearly riddled with unanswered questions at present. After all, it is not yet clear how W-CDMA will interoperate with cdma2000. While starting with the aim of a harmonised global standard, and being evolved from existing systems, the number of protocol combinations that have to be accommodated to allow roaming across the generations has multiplied. Granted that 2.5G networks have yet to be worn in, and licenses have yet to be awarded - but as Neil Montefiore at M1 says, commenting on the experience of implementing WAP roaming with C&W HKT, "we had more problems than we thought we'd have. I'm hoping that before too long everything will work easier together."
THE NEED TO AGREE
While GSM operators negotiating roaming agreements work within guidelines laid out by the GSM Association, there is no such framework for agreements between operators running, for example, ANSI-41 and GSM networks - they haven 't had to talk to each other in the past. As Aidan Dillon, wireless mobility manager at Logica's mobile networks division, points out, "operators will be dealing with technologies they haven't dealt with before - the GSM world now has to have an understanding of ANSI 41 and vice versa."
BACKWARDS COMPATIBILITY
Backwards compatibility is included in the recently published Release 99 specifications of the 3rd Generation Partnership Project (3GPP) - the body established to prepare technical specifications for a third generation mobile system based on evolved GSM core networks. The handover between UMTS and GSM is not yet fully specified, although it's anticipated that this will be rectified in the coming months. The 3GPP has not yet decided whether there is a market need for handover to be developed for other 3G systems. Similarly, backwards compatibility for the IP version of UMTS, which will be covered in its next set of specifications - Release 2000 - has not yet been tackled.
totaltele.com |
