WiMAX summit: 'Standards-plus' could harm 802.16 roadmap [printer-friendly]

Original URL: http://www.theregister.co.uk/2005/04/11/wimax_summit/

WiMAX summit: 'Standards-plus' could harm 802.16 roadmap

By Wireless Watch (peter at rethinkresearch.biz)

Published Monday 11th April 2005 13:22 GMT

This week saw the WiMAX community gathered in force at the WiMAX Summit in Paris, France. It quickly emerged that the issue preoccupying both vendors and potential operators is the road to mobility and exactly how the transition to the forthcoming 802.l6e mobile standard will be achieved. With a key WiMAX Forum meeting to be held in the coming week in Spain, and 802.16e set to be ratified this year, it is essential to the uptake of the platform that the route to mobility is clarified as soon as possible.

All agree that 802.16 will be the platform with which WiMAX hits the big time. Most of the equipment majors are merely licensing fixed 802.16d (now renamed 802.16-2004) gear, while focusing their own development efforts on 'e'. That means that the chances for chipmakers to net the big OEM deals - with Alcatel, Nortel and the others - rely on the mobile standard. But there are two basic schools of thought among the chipmakers and their licensees as to their strategy in the interim.

The two WiMAX roadmaps

One is that there is a period of at least two years before 802.16e achieves volume, and that the upgrade path will be complex. That means the priority is to make 802.16-2004 as impressive as possible in order to drive short term sales and increase confidence in WiMAX. This will mean creating a so-called 'd+' technology that goes beyond the basic stipulations of the fixed standard, with a focus on aspects such as quality of service for voice and video, and portability with consumer grade subscriber equipment.

The other view is that the market needs to move to mobility more rapidly, by offering pre-standard networks that provide most of the functionality promised for 'e'. This strategy rests on the belief - or hope - that the mobile standard will come to market rapidly and that the leap from its predecessor will be a simple one.

Both points of view reflect that 802.16-2004, in its vanilla form, is fairly basic and will not, out of the box, deliver all the services and capabilities that have been marketed on its behalf. Either 'd+' or 'e' will be required to fill those gaps and make WiMAX an attractive platform outside the traditional broadband wireless niches such as backhaul and rural access, but these will bring the risks of non-standard extensions that kill interoperability, a trend that is already afflicting WiFi.

The main decision that will affect how painless is the shift from 802.16-2004 to 802.16e rests on the choice physical layer (PHY). The more complex the upgrade, the more important it is to create an attractive first generation platform - whether 'd+' or pre-standard mobile - that will justify investment in its own right and deliver returns before the upgrade process becomes necessary.

The PHY debate

The original IEEE/Forum plan was to use the same PHY for 'e' as for the fixed specification - OFDM 256 FFT. This would enable a simple, often software-based upgrade and would support interoperability of fixed and mobile base stations and customer premises equipment (CPE). The agreement last year to incorporate the Korean WiBro technology into the 802.16e process - a highly positive move in terms of international harmonization and technical knowhow - did however introduce the option of using the PHY favored by Samsung and LG. This is Scalable OFDMA, which claims better support for mobility by supporting sub-channelization and variation of the size of FFT according to bandwidth available. Despite its advantages, SOFDMA equipment would require hardware upgrades to base stations and would not be interoperable with legacy 802.16-2004 equipment, even in the same frequency band.

This would be bad news for vendors that have put significant investment already in to fixed WiMAX, and for those that believe the big customers are waiting for mobility - and may not make the WiMAX move at all until they see an easy route to 802.16e. Some argue that the OFDM 256 PHY must be preserved in order to make the 802.16-2004 a valuable platform. WaveSat, for instance - the first chipmaker to get 802.16 compliant silicon to market - is adamant that sub-channelization can be supported in OFDM 256 and that this would be as strong a basis for mobility as SOFDMA.

This decision needs to be made, one way or the other, in the near future before operators will have a clear picture of how they can move towards mobile WiMAX. Until they do this, they will remain confused and frustrated and are unlikely to make major investments without a clear route forward.

The 'd+' route

Some, of course, especially in the wireline community, are interested in adopting any flavor of WiMAX as early as possible in order to expand broadband services into new markets. For these, the capabilities of 802.16-2004 are what matters and they already have high - some would say inflated - expectations of what standards certified gear will deliver when it ships in the autumn. Aware of the danger of disappointing early adopters at a time when WiMAX is already facing something of a backlash, chipmakers and vendors are racing to build as much functionality as possible into the platform.

French chip start-up Sequans is one example, with what it calls its 'd+' platform. The company is seeking to differentiate itself from the larger silicon vendors with a focus on some key operator criteria such as advanced quality of service, and on a highly integrated system-on-a-chip (SoC) that will be suitable for low cost micro base stations and small consumer devices. Head of marketing Ambroise Popper knows that the real volume will come with 802.16e and the big OEMs, and the company has an 'e' development project underway, but he believes there is a two-year opportunity for d+ devices delivering services such as voice over IP and video on demand. Sequans also supports wider channels than Intel, up to 28MHz, which will be important for backhaul applications.

WaveSat has a similar message. CEO Michel Guay says: "Don't wait for mobility, it will be 2-3 years. We want to offer a solution now with evolution to mobility" - though that evolution will require a hardware upgrade. WaveSat's primary focus is on getting fixed WiMAX into the mass market quickly. Although the major OEMs are the main customers this year, in 2006 he sees the original design manufacturers (ODMs), particularly in Taiwan, entering the sector and bringing down costs quickly.

Then the WiFi box makers will start to incorporate WiMAX. The critical development for this broad market will be support for the Mini PCI interface in WiMAX, something WaveSat promises by midyear and which will be important in the enterprise, supporting WiMAX networks with WiFi CPE. The first equipment maker to ship this combination is likely to be Alpha Networks of Taiwan, which will combine an Atheros 802.11 chip with a mini PCI WiMAX card, targeting hotzones.

To move towards mobility, the Canadian company will support sub-channelling this year, and clearly hopes that will enough, though it is committed to create chips for SOFDMA if the WiMAX Forum mandates it. It is following a more cautious route than Sequans and other towards a full SoC, claiming it is better to wait until the standard is more stable before incorporating the whole MAC on one chip.

Migration to mobility

Taking the opposite view is the UK's PicoChip, which is promising a software upgrade to 802.16e even if it takes the SOFDMA route. "We're 100 per cent sure we can software upgrade to 802.16e, although with SOFDMA it won't be interoperable with 802.16- 2004," said head of marketing Rupert Baines. The 'e' support will be a new software option for PicoChip's base station reference platform, which is heavily software defined, and both 'd' and 'e' will be accommodated within the same base station. The company believes its software approach is key to attracting some large OEMs, soon to be announced (its public customer is Airspan, which uses PicoChip for large base stations and Sequans for micro ones.)

PicoChip, like another start-up, TeleCIS, is expected also to take a prominent role in Korean WiBro, which could lead to early revenue opportunities since the Korean carriers are set to deploy the technology around the turn of the year.

Enhancements

Back in fixed wireless, there are several technologies that are being eyed by the chipmakers to improve basic WiMAX performance. As in WiFi, these include various antenna techniques to enhance data rates and range. Key are adaptive antenna systems, also known as beam forming, space-time coding and MIMO (Multiple In Multiple Out) arrays of smart antennas. Chipmakers are sure to implement some of these in order to improve performance and differentiate themselves, but unless they are included within the WiMAX specifications, they will cause interoperability problems between different vendors' equipment. And interoperability is critical. Indeed, as Redline's head of business development, Kevin Suitor, points out, it is the chief attraction of WiMAX before it achieves other promised benefits such as low cost CPE.

This week's Forum meeting in Malaga, Spain will, it is hoped, make significant progress on defining requirements for these three technologies, all currently optional modes. But even more important will be the decision on whether to incorporate these modes in the certification process, the only way to ensure interoperability and avoid the 'standards-plus' products and vendor lock-in that is being seen in the Wi-Fi community as suppliers incorporate MIMO and beam forming, before these technologies are supported in the upcoming 802.11n fast WiFi standard.

Of course, the more requirements that are included in certification, the more complex the process and the specification, and the more likelihood of further delays to shipment of approved equipment. However, this should still be preferable to hasty delivery of gear that does not meet service providers' needs. In the absence of sophisticated functionality within 802.16, operators that want to deliver services more advanced than basic internet access or backhaul will look to proprietary alternatives.

Early mobility

In the mobile-oriented world, some vendors believe operators should do this anyway, achieving mobility now with nonstandard networks, provided they can secure a strong guarantee from their supplier that there will be an upgrade to 802.16e, with its pricing and interoperability benefits, in a reasonable time span (two years or so). Navini and Clearwire's equipment subsidiary NextNet are both taking this approach.

Charles Riggle, VP of business development at NextNet, claims: "We believe NextNet's products are WIMAX ready even though there is no ratified standard - in terms of features and functionality." Because it is not forced to follow a standards timeline, NextNet says it can leap ahead in some key functionality, notably a laptop-based CPE. Its third generation architecture will be embedded in a PCMCIA card by mid-2006. The company will develop its standard 802.16 product with Intel.

The issue for these pre-WiMAX mobile companies may turn out to be, not whether they can deliver functionality rapidly enough, but whether they will have to take a step backwards in order to support the 'e' standard. There is considerable pressure on the suppliers, with their technology roadmaps, and on the Forum, in terms of the complexity of options it can support, to ensure that particular dilemma does not arise to confuse the market further. Firm decisions and clear strategies will be required in 2005 to prevent dangerous fragmentation of the WiMAX market from 2006.

Wireless Watch is published by Rethink Research, a London-based IT publishing and consulting firm. This weekly newsletter delivers in-depth analysis and market research of mobile and wireless for business. Subscription details are here (http://www.rethinkresearch.biz/subscribe.asp).