Original URL: http://www.theregister.co.uk/2004/07/16/ieee_groups_slug_it_out/

IEEE groups fight for control of key standards

Pace quickens

By Wireless Watch

Posted in Broadband, 16th July 2004 12:58 GMT

Analysis This week’s IEEE summit highlighted the breakneck pace of change that is driving innovation in wireless, but also threatening to break its standards process apart. Political wars rage in areas like UltraWideBand and fast Wi-Fi, but more fundamental debates are taking place over how different specifications should coexist and which territory they should occupy. As Wi-Fi reaches up to WiMAX’ range and WiMAX aims for the mobility of 802.20, the most important IEEE group of all may be 802.22, looking at the cognitive radio that will enable devices to use all three and to take advantage of proposed opening of US television spectrum.

This week’s IEEE meeting in Portland, Oregon is shaping up to be one of the most lively in years, with a packed agenda that highlights both the current breakneck pace of change in wireless networking, and many of the splits that could slow this pace in the near future.

As well as the obligatory vote on the UltraWideBand-based 802.15.3a standard proposal, in which the two technologies – Motorola’s DS-UWB and Intel/TI’s Multiband OFDM – will fight it out once more, there are new debates raging that may prove just as intractable. Like the UWB schism - which threatens to drive both camps to establish de facto standards, diluting the role of the IEEE in the wireless future – splits over fast Wi-Fi, mobile WiMAX and use of broadcasting spectrum for broadband wireless could all have a highly negative effect on the progress towards a multi-network, standards-based wireless world.

The fight for broadcast spectrum

A more unexpected battle, but one that shows where the ambitions of the WiMAX camp are heading, broke out early in the IEEE summit, over television broadband. Proposals to allow sharing of underused US TV spectrum by wireless devices could open a whole new range of opportunities for WiMAX, and the 802.16 group and its greatest cheerleader Intel are keen to take the driving seat.

These ambitions clashed with IEEE plans to set up a separate body, called 802.22, to look at a standard for the peculiar requirements of broadband wireless in TV spectrum, and specifically at a cognitive radio that would be able to detect when a wireless device might interfere with an incumbent broadcaster’s signal and back off. This group would be a spin-off of the existing 802.18, also called the Radio Regulatory TAG, which exists to provide the 802 wireless working groups with regulatory expertise.

This plan was denounced by the 802.16 committee chairman, Roger Marks, who claimed its work was already covered by the remit of the WiMAX work. "I don't see how this is in anyway different than what we do in 802.16," said 802.16 members. A resolution was passed that would submit comments to the Executive Committee, explaining why the TV-Broadband standards work should be incorporated into the 802.16 standards group and not spun off into a separate working group like 802.22.

The debate may seem obscure, but a look at the factors behind it reveal why it is actually of crucial importance to companies that want WiMAX to become the dominant technology for broadband wireless.

Last month, the Federal Communications Commission (FCC) issued a Notice of Proposed Rulemaking that would allow for unlicensed broadband wireless operations in underused television frequencies, in channels below 900MHz and in the 3GHz band. This proposal has been heavily supported by FCC chair Michael Powell and has been the subject of intensive lobbying by Intel, which sees the additional spectrum as a rich hunting ground for WiMAX, particularly in its mobile form.

The FCC proposes to insist that unlicensed devices in these bands should incorporate cognitive radios to identify unused channels. It will permit two types of broadband access – by low power personal gadgets such as Wi-Fi or WiMAX cards in laptop computers or home networks; and by higher power fixed wireless technology providing commercial broadband services.

As well as relieving the pressure on overcrowded unlicensed spectrum for Wi-Fi and WiMAX operators, the freeing up of these frequencies would also be important because signals in the lower TV band can travel longer distances and penetrate buildings and trees more easily than they can in the current 2.4GHz and 5GHz bands.

Intel’s CTO, Pat Gelsinger, a vocal campaigner for more liberal approaches to spectrum suitable for Wi-Fi and WiMAX, commented:

"For more than half a century, vacant TV channels (which represent some of the most valuable spectrum available) have been underutilized. Releasing this spectrum for unlicensed use will help foster new technologies, create opportunities for business and bring exciting new products to consumers."

The move will be a driver to accelerate work on WiMAX profiles that will allow the technology to work in the 3GHz band and the lower frequencies. Although all the focus of recent 802.16 work has been on bands between 2GHz and 11GHz, the WiMAX Forum said earlier this year that it would also look at profiles for low frequencies, clearly with the potential FCC changes in mind.

WiMAX is flexible in its channel sizes and can use the 6MHz width of the TV channels. Even with these narrower channels, signal range below 900MHz could be three times that in 2.4GHz, reducing the number of base stations required well below 3G’s requirements, and so making mobile WiMAX clouds an even stronger proposition against cellular.

Slow progress to mobility

All this shows why Intel and the WiMAX community are so keen to establish their influence over the opening sub-900MHz market and ensure that standards and devices that can take advantage of this goldmine of new spectrum are geared to the vision of ubiquitous fixed and mobile 802.16.

However, the proposed mobile variant, 802.16e, has a tough job to hit its end of year deadline for defining a standard for the core WiMAX frequencies, let alone any new ones. A total of 122 contributions was submitted for the 802.16e working group this week and few of these have harmonized, though they are expected to boil down to two core groups of proposals.

One of these will have heavy input from Navini Networks, the developer of highly mobile broadband wireless technology, which joined WiMAX Forum earlier this year, leaving behind its former favored standards body, 802.20 or Mobile-Fi. Navini brings a very valuable source of expertise in high speed handoff to the WiMAX camp, which was originally focused on limited portability, leaving true mobility to its 802.20 cousin. Now the remit of 802.16e has broadened to cover both portability and mobility, and there is increasing pressure for Mobile-Fi to be integrated into its work rather than pursuing a separate and increasingly overlapping agenda. This view is strengthened as more Mobile-Fi backers join WiMAX, including Motorola and, potentially, Cisco.

There is something of a split between the IEEE 802.16 committee and the WiMAX Forum industry group over coexistence with another overlapping standard, the South Korean Wi-Bro, which was largely developed by Samsung, research body ETRI and several Korean operators. Marks said at this week’s meeting that he welcomed the large Korean involvement in 802.16e and that the expertise of members from that country had “contributed substantially” to the emerging standard.

This conflicts with the views stated by various Intel representatives, including WiMAX chairman Ron Resnick – the chipmaker has opposed Samsung’s claim that Wi-Bro could form the basis of 802.16e, reducing time to standards, since the Korean technology is already developed and will be deployed by SKT next year. Marks seems to be taking the view that the overlapping specifications should be harmonized under one standard, even if this is then deployed in different ways for different markets – something to which WiMAX, with its broad set of potential profiles, lends itself. For instance, Wi-Bro uses 2,048 OFDM carriers, a variation that is supported in 802.16 by the Multiple Access Mode, though the 802.16e work is mainly focused on 1,024 carriers. Marks is sceptical that the goal of Wi-Bro, to deliver 1Mbps data rates at high speed, is really needed in the market, a view that may reflect the differences between US and Korean users.

In-car 802.11p standards

While the 802.16e project is in danger of being delayed in its progress to high speed mobility by these politics, its thunder in the vehicle market could be stolen by Wi-Fi. The next extension to the 802.11 standards family to come under consideration by the IEEE will be 802.11p, which concerns Wi-Fi operations at high speed and is seen as an important enabler of in-car networks. The proposed 802.11p protocol would deliver 6Mbps average rates over distances of around 1,000 feet while travelling at high speed, making it competitive with other mobile broadband technologies such as Flarion Flash-OFDM or mobile WiMAX. It expands on conventional 802.11 to allow for provisions that are specifically useful to automobiles - a more advanced hand-off scheme, high mobility, enhanced security, identification, ad hoc peer-to-peer authentication and communications in the automotive- allocated 5.9GHz spectrum.

802.11p will be used as the groundwork for DSRC (Dedicated Short Range Communications), a US Department of Transportation project – which will be emulated elsewhere - looking at vehicle- based communication networks, particularly for applications such as toll collection, vehicle safety services, and commerce transactions via cars. The ultimate vision is a nationwide network that enables communications between vehicles and roadside access points or other vehicles.

The 802.11p technology will come before the executive committee of the IEEE this week and working groups are conducting further studies on interoperability between access points and cars. The work builds on its predecessor, ASTN a2213-O3.

“Prototypes are under construction right now,” said Lee Armstrong, chair of the 802.11p working group, who expects auto makers to deploy the technology in high end vehicles in 2007-8, and says most of the big names are already testing the connections. By building Wi-Fi access points along the nation’s highways— which ABI estimates could see initial expenditures in the $1bn range in the US—and installing receiver chips inside automobiles, consumers and safety personnel will be able to receive high speed internet access that supports such applications as real time traffic updates and video streaming.

This could eat into a valuable market for both cellular and broadband wireless equipment makers and operators. Although the current default method of communication to the car, cellular, will exist alongside 802.11p for at least another six years, it is possible that carmakers will hold back on investing in next generation cellular telematics technology, which has been held up as a major new revenue stream for this industry.

Likewise, 802.11p will take some of the gloss of the claims of Flarion and other 802.20 backers, as well as companies looking at high speed WiMAX, that their technologies will be differentiated by their suitability for vehicles. The government support for 802.11p will make it the mainstream solution and alternatives, even if they are more powerful, will find it hard to get a look-in.

New divisions in 802.11n:

Also in Wi-Fi, further divisions have arisen in the quest for a fast version of the standard. We have already outlined the division over the technological approach that should underpin the next extension to Wi-Fi, 802.11n, which will specify networks operating at over 100Mbps (see Wireless Watch issue 69). Now, challenging the two main groupings, led respectively by Atheros and Airgo, is chipmaker Agere, which has thrown the cat among the pigeons, claiming the real goal of the standard should be 500Mbps data rates.

The company says its proposal combines MIMO (multiple in multiple out) smart antennas with single 20MHz or double 40MHz channel widths. The full 500Mbps raw throughput is achieved by using the maximum number of antennas – four at each transmitter and receivery – and the wider channels. The technology would offer backwards compatibility with 802.11b/g and would also support two-antenna configurations for lower cost products at lower speeds.

The company has been working on wireless MIMO technology for some years and in 2002 demonstrated a 162Mbps three transmitter/three receiver MIMO system.

Agere would not provide details of when working silicon would be available to demonstrate its proposed 500Mbps 802.11n capability. It said the flexibility to support single or double-width channels not only made it suitable for all countries – Japan specifies 20MHz channels for WLans, for instance – but also allowed for a configuration in 2.4GHz. While 5GHz would be required for the full 500Mbps, the lower frequency support would allow for coexistence with 802.11b/g.

Another feature of the submission is a standardized frame aggregation for both single and multiple destinations, to improve network efficiency and interoperability. Frame aggregation, - which merges several frames together in a single packet, - is important for streaming applications including voice over IP and multimedia content.

The road to harmonization

All these disputes are to be expected in a wireless world that is so newly formed and where all the technologies and standards, and many of the players, are immature. Some of them, like the 802.11n and 802.15.3a splits, do raise important issues of functionality, but are primarily about vendor bids for control and royalties.

Others, like the arguments over whether to unify 802.16, 802.20 and Wi-Bro, while political, are also centered on fundamental debates over the role of the various standards in the real world. Should there be separate standards for portable, metro area wireless and highly mobile networks, or should these be merely different variations of the same base?

There are technological and market arguments on both sides, even if the final outcome is likely to be decided by the needs of the big vendors, like Intel, to support as broad and powerful a standard as possible. What the current IEEE meeting also highlights is the dilemma created by the very speed of wireless technology development.

Rapid uptake of Wi-Fi has attracted hundreds of suppliers, all seeking to survive, and to maintain margins, by stretching the technology a step further and gaining differentiation from their rivals. This cycle of price wars and new product developments is more rapid than we have seen anywhere before in the communications sector and, while it fosters innovation, it also creates huge problems for standards bodies and therefore for operators’ and enterprises’ longer term planning.

If Wi-Fi can move out of its LAN space and achieve 100 miles per hour mobility, 500Mbps data rates and extended range, why do we need WiMAX? If WiMAX can reach beyond its metro area remit and become fully mobile, what is the point of 802.20? And that’s not to mention the cellular technologies still determined to hold on to their massive market lead in wireless communications.

The answer, of course, comes back to harmonization, effective integration and the much vaunted cognitive radio, all of which will make the choice of standards less critical and will allow devices to take advantage of the different benefits of all those available. Reduce the list down to 802.11, 802.16, 802.15 and cellular, and allow users to access all of them on an always best connected basis, and we are coming close to the ubiquitous wireless world touted by the equipment makers and Michael Powell.

In this picture, the cognitive radio becomes the most vital wireless development of the coming few years , and the mooted 802.22 committee perhaps the most important of all within the IEEE. Which finally makes it very clear why the WiMAX brigade are so keen to bring that effort under their auspices.

Upcoming IEEE standards

The next Wi-Fi standards in line to be ratified are 802.11e for quality of service, later this year, followed in the 1-2 year timeframe by 802.11n for 108Mbps speeds, 802.11r for fast hand-off between access points, and 802.11s for mesh networking.

Further down the track is 802.11p, and further taskforces are mooted for performance prediction for testing (likely to be 802.11t) and interworking with external networks. It is also possible that a standing committee will be formed to look at security issues on an ongoing basis, especially to ensure that these new extensions do not involve new security problems, and there is also the 802.11m standing committee, for standards maintenance.

For 802.16, the next priorities are the mobile 802.16e, tabled for late this year, and an 802.16f extension addressing mesh using directional antennas, slated for late 2005.

For personal area networks, the fight drags on to choose the base for the 802.15.3a standard, a higher speed, UWB-based extension to 802.15.3 or WiMedia. While 802.15.3 addresses high data rate, short range networks, its cousin 802.15.4 or ZigBee looks at low data rate, low power equivalents. Its current major project is an ‘a’ extension of its own, also based on UWB.

Copyright © 2004, Wireless Watch

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.

Related stories

Wi-Fi group intros standards support stamp
Wi-Fi group to update WLAN spec
Motorola and MBOA split on UWB
March is crunch time for Mobile-Fi