“WiMAX is a good technology,” commented Matts Sporre, mobile director of TeliaSonera, “But what’s it good FOR?”

Sporre was speaking in September at Informa’s WiMAX 2006 conference in Prague. An inspired choice of speaker. TeliaSonera runs WiFi hotspots as well as fixed and cellular networks. They operate in the Nordic and Baltic countries, in Russian and Eurasian markets as well as in Spain – a wide range of geographies and demographics encompassing both developed and developing economies. Their interests include hotspots, broadband access, backhaul and mobile access. If WiMAX can’t fit somewhere within TeliaSonera’s technology roadmap then it’s got real problems.

Hotspots
“Wherever you see WiFi today, expect to see WiMAX tomorrow,” vowed Andrzej Galan, business development manager of Intel Technology Poland. Few would dispute that promise. Intel is moving inexorably towards its goal of ubiquitous availability of WiMAX in devices – hoping to drive the market to a tipping point just as it did with WiFi. Hotspot access is clearly a strong prospect for WiMAX deployment.

TeliaSonera’s HomeRun service has over 700 hotspots and roaming agreements with 3500 hotspots in 16 countries, announced Sporre. WiFi is therefore already widely deployed and is inter-working with UMTS/HSPA. Why would they want to change? To reduce costs perhaps. But hotspot costs are dominated by floor space and electricity, WiMAX won’t change that. Perhaps because WiMAX will be built into laptops? But WiFi is already built into laptops. “We’ll continue to use WiFi as the technology for hotspot access in the medium term,” stated Sporre.

In the longer term WiMAX is not the only candidate for a next generation hotspot technology, argued Sporre; it will compete with HSPA+ and UMTS LTE. Technology built into laptops will be the trigger for the next generation. “This business is terminal driven; technology needs to be built in to reach a wide customer base.”

Over to you, Intel.

Broadband access
Broadband wireless access was the original motivation for WiMAX. But from TeliaSonera’s perspective WiMAX can’t compete with fibre or with xDSL in urban, suburban or even rural areas if decent copper exists. Where it could compete is in rural areas without good copper. In some countries these areas are large; creating an opportunity for a radio based broadband technology.

Whether you believe that technology should be of the fixed WiMAX (802.16-2004, commonly known as “d”) or mobile WiMAX (802.16e-2005, referred to as “e”) variety is largely a matter of religious persuasion. The fixed WiMAX variety can do nomadic which is good for a DSL alternative. Certified equipment now exists for fixed. That’s also good but don’t get too excited – certified does not yet mean interoperable. In fact certification may never imply interoperability. This is the WiMAX world remember, where words take on elastic and ambiguous meanings. The mobile WiMAX variety is also good for a DSL alternative as it can of course do fixed and nomadic as well – it can’t actually do mobile yet by the way but that will come. Probably, just don’t ask when. All clear now?

Devotees of fixed WiMAX or “d” seem to be dwindling, increasingly confined to vendors anxious to express continued support for existing customers. Despite the obfuscating nature of the nomenclature, most people have now understood that fixed and mobile WiMAX are incompatible. “The d/e standards are entirely separate technologies and there is no straightforward transition from one to the other,” confirms John Munnery, coo of WiMAX start-up Max Telecom in Bulgaria.

Max Telecom is focusing initially on data connectivity, offering VoIP, DSL equivalent and VPNs. But they are installing “e”. “All new R&D investment seems to be going into 16e,” noted Munnery, “In the future 16e will dominate the WiMAX market, and even companies with fixed only licences will find they will have to consider purchasing 16e equipment as scale economies drive down prices.”

Not everyone agrees. In-building coverage is critical for broadband access and this could be a problem for “e”, particularly at 3.5GHz. New applications and appliances are changing consumer behaviour reported Jeremy Nel of the broadband wireless access operator Irish Broadband. Poor in-building penetration may be acceptable for mobile handsets but not for laptops – people won’t carry their laptops to a different place in a building to get coverage.

“Operators need to choose one or the other, or if already committed to “d” will have to run a parallel operation,” concludes Munnery. Talk of migration from “d” to “e” is nonsense according to Munnery: “The network topologies are completely different and “upgrading” would involve replacing all the base stations as well as all the CPE equipment.”

Perhaps not all the CPE equipment. Intel’s new Rosedale 2 chipset will have the capability to support both “d” and “e” flavours. “Service providers will be able to deploy modems based on Rosedale 2 for use in fixed applications and then evolve their WiMAX networks from 802.16-2004 to 802.16e-2005 with a software update,” is the Intel party line.

We’ll wait and see was the reaction in Prague. Rosedale 2 may include the capability for both fixed and mobile WiMAX but upgradeability is a different matter.

Over to you, Intel.

Backhaul
There’s simply no need for WiMAX backhaul solutions if you already have good radio links and fibre infrastructure, argued TeliaSonera’s Sporre. But if the transport networks are no good or you are extending coverage into remote areas then opportunities could exist for WiMAX. The Eurasian markets in TeliaSonera’s portfolio have poor transmission in some places, admitted Sporre.

Fixed WiMAX backhaul solutions deploy directional antennas in point-to-point, line-of-sight configurations in which all power is delivered to a single fixed connection. They work just fine. But their advantages compared with conventional microwave links are somewhat marginal. Why should a service provider bother with such a solution?

Dean Chang, Chair of the Service Provider Working Group at the WiMAX Forum, had a suggestion. Deploy “d” today and then overlay with “e” tomorrow, retaining “d” for backhaul. That gives you a fixed to mobile migration path, he argued.

Max Telecom’s Munnery was not impressed. You could end up using a third of your spectrum allocation for backhaul, he commented. Andy McKinnon, Principal WiMAX EMEA, Motorola, noted that fixed and mobile deployments impose different spectrum requirements. “Radio performance and in-building coverage are very different in “d” and “e”,” said McKinnon, “deploying fixed, portable and mobile solutions in a single spectrum allocation is not easy.”

Motorola do not make “d”, all their WiMAX equipment is based on “e”.

Mobile access
“Certified mobile WiMAX products are not yet on the market,” observed Sporre, “and their performance will be similar to HSDPA.” So UMTS operators will naturally opt for HSPA upgrades – mobile WiMAX is just too late. For operators without a 3G licence, however, WiMAX can be an alternative.

WiMAX can be an alternative for operators who missed out on 3G licences if they can get hold of suitable spectrum. If they have to use 3.5GHz spectrum, however, then delivering nationwide coverage becomes a major problem. If they have to use 3.5GHz then delivering mobility also becomes a major problem.

So at first sight it’s a bit surprising to find what’s billed as the world’s first nationwide WiMAX deployment – Wateen Telecom in Pakistan – is using 16e mobile WiMAX equipment at 3.5GHz. On second sight it’s not so surprising when you consider the network is heavily funded by Motorola, who only produce “e”.

Wateen is deploying a fixed network with initial rollout to outdoor CPE. Probably not because their licence only covers fixed usage – Wateen adopt a rather cavalier attitude to regulatory constraints common within sections of the WiMAX community – but more because Wateen’s owners, Warid Telecom, also run a GSM network in Pakistan.

The other proud owner of both cellular and WiMAX networks of course is Sprint-Nextel, although they have the advantage of massive spectrum holdings at 2.5GHz. Sprint-Nextel is upgrading its existing EV-DO network to Rev A whilst simultaneously building out a WiMAX network – with strong ‘backing’ from suppliers Intel, Motorola and Samsung. Sprint-Nextel has to begin offering services in the 2.5GHz band to at least 15mn Americans by 2009 and to 15mn more by 2011, or risk losing rights to the spectrum. And they’ve chosen to deploy a new technology without any existing equipment. No pressure there then.

Heavy vendor financing is widely assumed to be a feature of the Sprint-Nextel WiMAX deployment. That may or may not be true – it’s neither been confirmed nor denied – but heavy vendor financing certainly is a feature of Clearwire’s rural US WiMAX deployments at 2.5GHz. Having scooped up US$600mn from Intel and US$300mn from Motorola, Craig McCaw’s Clearwire is on a roll, combining myriad regional licences together to generate coherent coverage. It worked before for McCaw and could be working again. It may even emerge in Europe where Clearwire is quietly amassing licences and already has ‘pre-WiMAX’ operations in Belgium, Denmark, Ireland, Spain and Poland.

These developments continue to be contrasted with mobile systems such as 3G cellular. But such comparisons could be flawed. WiMax is a set of access technologies, not a network solution. Technologies such as WiFi and WiMAX are IP-only ‘access bridges’ providing connection to the fixed network. Although the WiMax Forum is working on a network architecture framework for WiMax a complete network solution including full mobility and roaming is a long way off.

Maybe it will never come. Maybe people are misinterpreting what WiMAX is all about. WiMAX is about making computing devices communicate; it’s really a consumer electronics proposition. Cellular on the other hand is more about making communication devices compute. Randall Schwartz, principal consultant with the consulting group WiMAX 20/20, provided the insight: “Development of CPEs will pace the market and determine which applications will be successful.”

Despite the Sprint-Nextel decision the mood in Prague was less than euphoric. Clear signs of relief were evident as WiMAX emerges from the hype of the past. But relief tinged with nervousness about the gritty reality of the future. That future still holds promise but remains uncertain.

A somewhat unexpected consensus emerged from the speakers in Prague: Without low cost, off the shelf consumer devices, WiMAX is dead. Andy McKinnon put it most succinctly; “We’re all reliant on the Intel marketing engine.”

Over to you, Intel.
Stuart Sharrock