LTE spectrum: Done Roaming

The allocation and usage of radio spectrum, that very scarce and valuable resource, varies wildly from market to market. And while national regulators have gravitated toward certain spectrum plans, they have not done so in sufficient volume that any of these plans could be considered properly international.

But with LTE deployments well under way in many mature markets as well as a growing number of emerging ones, international and domestic roaming issues are being brought to the fore. The big problem is that with individual countries at the whim of their own respective market forces, harmonisation even within a large country like the USA is a tall order.

Over the last several years, spectrum in the 700MHz band has been made available around the world through the transition to digital TV, as frequencies previously used for UHF television broadcasts were refarmed and re-auctioned. Indeed, there were high hopes for the harmonisation of the 700MHz band on an international roaming level for LTE, with plans afoot to make use of the same 700MHz spectrum in the US, Asia Pacific and Europe. As it stands, 700MHz is the third most widely used band for LTE deployments after 2.6GHz and 1800MHz (currently the clear leader). But the 700MHz band is actually made up of frequencies from the 600 to the 800 MHz range and each region’s demarcation is slightly different. And as regional plans have developed, it’s becoming clear that any hopes of using 700MHz for roaming internationally will likely be scuppered by fragmentation and disjointed strategies.

The US was the first region to move on the 700MHz frequency (actually consisting of spectrum in the 698 to 806MHz bands) for cellular networking and in this case, the 700MHz band is actually made up of two separate bands: the lower band at 698-746MHz and the upper band at 746-806MHz. Only the A, B, and C blocks are paired in the lower band (leaving two unpaired blocks) but the upper is divided into five paired blocks. When the band plans were originally created, the A, B and lower C blocks of spectrum were intended to share a common band class, known as Band 12, and as a result, to share radio components. But, citing interference concerns with broadcast TV in the A block, AT&T successfully lobbied for the creation of a separate band, Band 17, containing only the B and lower C block. This is effectively a proprietary band and got AT&T out of using A Block, Band 12 chips in its devices. But the absence of a heavyweight carrier using Band 12 (blocks A, B and lower C) gave chipset manufacturers little incentive to develop terminal radios for this specific band. At least this is the argument put forward by several of America’s regional operators. About 12 months ago Cavalier Wireless, C Spire, Continuum 700, King Street Wireless, MetroPCS, US Cellular, and Vulcan Wireless filed a petition with the FCC claiming as much, but also asserting that AT&T’s real motive was to squeeze smaller players out of the LTE market.

Taken in conjunction with Verizon’s approach— which with a good chunk of the upper band in its portfolio sold off some of its A block spectrum and concentrated on the upper C block where it almost has enough spectrum to roll out a nationwide network in the 700MHz band—it certainly looks like the smaller players will have a hard time getting in on the LTE action as their networks will not be interoperable with the largest operators in the market.

Many of these smaller carriers in the US, which operate on a regional level, only acquired A block spectrum. And since AT&T effectively marginalised the A block, these smaller players claim to have struggled to get hold of devices. After all, manufacturers are mainly interested in large volume production, but also cannot sell affordable devices if there is not the scale to mass produce them. Many of these same rural carriers are calling for an extension to build-out deadlines as a result.

As a company with significant scale, Verizon Wireless, has no such problems. The operator is currently running the world’s largest LTE network, boasting service availability in 491 US markets covering 287 million POPs. At present Verizon offers LTE coverage over more than 95 per cent of its 3G deployment area and expects to match its 3G network with LTE coverage by the end of the second quarter and at year end at the very latest.

The operator has also partnered with some of the smaller, rural players through its LTE in Rural America programme. A Verizon spokesman says the company is pursuing an aggressive growth plan involving rural partnerships where it doesn’t have LTE coverage. “Over the past couple of years we have worked to bring LTE beyond major cities with this programme. We work with rural communications companies to build and operate a 4G network in their markets using their tower and backhaul assets and our core equipment and premium 700MHz spectrum assets,” he says. But while this gives regional players and opportunity to participate in LTE, it doesn’t help the suspicion that block A spectrum was sold as a lemon.

The LTE in Rural America scheme is currently in action in 49 states and will soon bring service into play in Alaska too. A total of 20 rural carriers are involved in the programme, covering 2.8 million people in rural communities and 179,000 square miles in 14 states. Eight of these players have already launched commercial LTE and Verizon is in active negotiations with others.

Stefan Zehle, CEO of spectrum consultancy Coleago Consulting, maintains that the 700MHz plan in the US is extremely inefficient. “It’s not a harmonised band, it’s effectively a proprietary band. The B and C blocks are effectively proprietary to AT&T and the upper part of the band is proprietary to Verizon,” he says.

“Normally in spectrum management you make sure you set standards in such a way that means anyone can use the system. But if the rest of the world had followed the US example they would have to follow Verizon or AT&T in handset availability and that’s not a great place to be,” he adds. Given that the rest of the world is some way behind the US in terms of 700MHz reallocation, if the curious US model was adopted globally, the two biggest US carriers would effectively be in charge of the direction of the global LTE handset portfolio.

But while the US model is unlikely to be adopted by any other region, it does look as though other regions might be pursuing somewhat separatist strategies. Many Asia-Pacific countries have already committed to the Asia Pacific Telecommunity (APT) plan, which takes 703–748MHz and pairs it with 758–803MHz. Australia, New Zealand, India, Japan, Korea, Papua New Guinea have now either formally adopted or signalled their intent to use the APT700MHz spectrum band for LTE services. Indonesia and Malaysia are yet to commit, though, while China—ever keen to plough its own furrow—has indicated that it aims to introduce TDD technology in the 700MHz band.

An allocation of 2x45MHz in APT is actually fairly wide for a lower band, so in markets which have three operators they could each be given 2x15MHz for very efficient usage. Even four operators could be well accommodated. Coleago’s Zehle compares this to the European 800MHz band plan which only had 2x30MHz swathes in Germany and France, with the result that one operator in each country was left out in the cold that couldn’t get any spectrum.

“Having a 700MHz plan set out according to the APT plan really gives much flexibility in terms of managing competition and avoiding issues we have in the US like the A block being useless. It really makes more efficient use of lower band spectrum for LTE,” Zehle says. Indeed, the APT plan proved attractive enough for many countries in Latin America to adopt. In February Chile’s telecoms regulator, Subtel, outlined plans to auction the 700MHz spectrum band for LTE services, along similar lines to the APT band plan.

Under the Subtel plan, frequencies in the 703MHz-748MHz range for uplink are paired with those in the 758MHz-803MHz band for downlink. With Subtel’s announcement, Chile follows in the footsteps of other nations in Latin America, including Argentina, Bolivia, Colombia, Costa Rica, Ecuador, Mexico, Panama, Uruguay, and Brazil, which have also opted in to the APT 4G scheme. Brazil’s Minister of Communications, Paulo Bernardo has officially sanctioned the use of the 700MHz band for mobile broadband services, and regulator Anatel has begun formulating how to define and allocate the frequency band for LTE use.

The switch from analogue to digital TV across Latin America will not be completed for a number of years, but Brazil’s government is looking to accelerate plans to switch off analogue signals in some parts of the country to July 2016.

“Brazil’s decision means that the APT eco-system is gaining the scale which confirms it as a mainstream solution for LTE deployment,” Zehle says. “This means the 700MHz APT band plan may appear in chipsets and more devices earlier rather than later.”

With Latin America and Asia Pacific creating so much momentum behind it, APT700 should have emerged as a promising international spectrum band. But with the North American and rest of the world LTE ecosystems diverging, device manufacturers will have to produce different iterations of their devices for Latin America and also cater to roaming bands between Latin America and Asia, no doubt hoping that the APT proposal gets adopted across the whole of the Asia Pacific region and perhaps more widely across Latin America. And despite regional enthusiasm for the APT model, it is just as incompatible with the European system as its North American cousin.

So, what about the Region 1 model incorporating Europe, Russia, the Middle East and Africa? A decision at the World Radiocommunication Conference in 2012 (WRC-12) created the possibility to allocate the 694–790MHz frequency band (the European/Region 1 ‘700MHz’ band) after WRC-15 in 2015 for mobile service. Despite the agreement not being set in stone, adoption is going ahead early in Africa where it has proved difficult to free up the 800MHz band that is already widely used for cellular services in Europe.

But since April, industry body the GSM Association has been lobbying for Europe, the Middle East and Africa to adopt an alternative 2x30MHz band plan for the 700MHz band, consisting of 703–733MHz (uplink) paired with 758–788MHz (downlink). This band plan is based on the reuse of the lower duplexer of the APT band plan (2x30MHz from the APT 2x45MHz) and is intended to promote greater interoperability.

The proposal paves the way for what the GSMA calls “a virtually global band with the APT band plan,” which “will offer the greatest benefit for consumers and mobile operators.” In its proposal, the GSMA recommends that the regulatory and technical conditions for the 700MHz band in EMEA be harmonised with the Asia Pacific band plan, “maximising economies of scale in device manufacturing and enabling interoperability and roaming.” But the spectrum allocation does rather differ to what has been proposed at the WRC. On the flip side, as well as less cross-border interference and more options for roaming, the GSMA plan would drive availability of a wider portfolio of devices, driven by a larger, international market, as device manufacturers can mass-produce less-complex devices that function in multiple countries on a single band.

“What’s nice about this plan is that it’s a little bit wider than last digital dividend for the 800MHZ band, which only had 2x30MHz, so possibly depending on the guard bands used you might get 2x40MHz out of that 700MHz allocation,” says Stefan Zehle. The benefit is being able to give four operators 2x10MHz each or have a slightly wider allocation. It’s good for competition in markets where there are more than three operators, which makes it important in Africa but also it’s in the lower band, which is good for coverage of areas with a lower population density, while still providing capacity in suburban areas because the allocation is quite chunky.

It gives the benefit of a cheap rebuild of geographic coverage and better in building penetration, which is key to the mobile broadband market. Indeed, some of the highest prices the industry has seen in terms of MHz per POP are for licences in urban areas, driven by in-building coverage and capacity needs.

Alan Hadden, chairman of the Global Suppliers Association (GSA), says the discussion is interesting because of the efforts to try and find a more efficient use of the 700MHz band. “There is this APT band plan which has been prepared and accepted and the FDD model is finding most support. It’s been adopted or studied by several operators in the Asia region, through to Australia and New Zealand, so there’s a lot of momentum around that.

“But the bigger story is what happens with Latin America. There’s been a widespread adoption of the APT configuration led by Mexico, Cofitel and others and you have the potential for a near global band offering good coverage at the 700MHz level but again sadly, North America will not be part of that as they’re not compatible,” Hadden says. Indeed, if the GSMA’s proposal is not accepted, the fragmented approach taken globally will put paid to use of the 700MHz band in any global roaming plans. The new Region 1 allocation is not in alignment with the upper end of the APT plan, and neither is Region 1’s 800MHz plan, while the US approach isn’t aligned with anything.

“700MHz will definitely not be a roaming band,” says Zehle. “The roaming bands we will see are 2.6GHz and 1800MHz. But ultimately the same thing will happen with LTE as is the case now. If you buy a quad-band handset in Europe you also have to have North American PCS in there and 850MHz in there too, and in time all these will be converted to LTE.” So that’s bad news for device manufacturers looking to reduce the number of radios in their handsets, and worse news for small carriers with more obscure spectrum portfolios which will struggle to get the most popular devices or get their own built with the benefits of economies of scale. “You can see the potential for scale, as there are really huge markets showing interest,” says Hadden, “APT 700 itself it is potentially very interesting. It’s like the 1800MHz model two years ago. But I don’t know what scenarios there will be for the industry other than relying on multiband devices,” he says.

“Not only is the US 700MHz plan inefficient in its overall use of spectrum, but matters are made worse because in part of the band the uplink and downlink are reversed,” says Zehle. “The device incompatibility issue with the US 700MHz band plan is that it splits the market into those who follow the AT&T device eco-system and those who follow the Verizon eco-system. These are mutually exclusive because the AT&T devices do not have the Verizon 700MHz blocks and vice versa. This means there is no harmonisation within the US.” Never mind the rest of the world. Other regions are showing signs of reluctance to standardise on the US spectrum plan for these reasons, as the situation would be exported to whichever country adopts the US 700MHz band strategy.

On the subject of exports, roaming is a big deal, especially for a carrier which at present sees 54 per cent of data traffic travelling over its LTE network. Verizon is aggressively seeking international partners to enable LTE roaming and expects to have the first of several agreements in place in 2014 so US travellers can take advantage of 4G networks overseas. “Canada will likely be first on that list,” says a spokesman, “but we still need to work out which technology and spectrum we will be using and so on.”

Of course, the industry is used to international roaming between continents and operators but history hides the nascent state of the LTE market, according to Ovum analyst, Steven Hartley. “The time it has taken to reach such a seemingly underwhelming point speaks volumes about the inherent complexities of roaming and the challenges that remain,” he said at the announcement of the world’s first LTE roaming deal.

The first carrier to tackle the cross border roaming issue was TeliaSonera, the Scandinavian company that was also first in the world to launch commercial LTE services. Since February, Danish LTE users have been able to roam on TeliaSonera’s LTE network but it has been duly noted that the fact that customers from a market that has had LTE for two years can roam to a network operated by the same company in a neighbouring country looks like nothing exciting.

According to Hartley, many challenges remain before LTE roaming becomes ubiquitous. On a global scale, there is spectrum fragmentation, as well as the operational perspective and bilateral nature of roaming deals that adds complexity. Yet there are murmurings from TeliaSonera that suggest an external roaming deal may be struck sometime this year. Yet when it comes to the USA, because of the way AT&T and Verizon have effectively divided the 700MHz band, European and Asian operators may still have to choose which US operator to align with.

The carriers that make the most noise are the ones likely to drive the ecosystem. North America has already carved its own path and Asia and Latin America look to set to travel a different route altogether. For Europe and the rest of the world a lot will hinge on the WRC conference in 2015, although there is evidence that Africa has already thrown its lot in with this model.

Adoption of the GSMA’s proposal could set Europe on a more harmonized footing with Asia and South America but all carriers seeking a roaming partner in the USA will still have to pick from Verizon or AT&T just as they have historically and all the time the world’s smaller carriers will have to toe the party line. LTE may have emerged as a global technology but it’s far from a global solution.