Standards and specs: A house divided

UWB's double standards

What happens when two coalitions within a standards committee come into conflict, and the dispute doesn't get resolved quickly? The ultrawideband technology standardization issue shows you.

Share:

Peter Seebach (crankyuser@seebs.plethora.net), Author, Freelance

Peter SeebachPeter Seebach has been using computers for years and is gradually becoming acclimated. He still doesn't know why mice need to be cleaned so often, though.



25 January 2005

As the previous Standards and specs column discussed, standardization is rooted in compromise -- and compromise can be hard. In some cases, competing interests leave some vendors in a position where they would rather fork the standard and launch their own sub-version, than continue working on resolving a conflict. Market pressures play a significant role in this. If customers demand a given piece of technology today, then you'll have a hard time explaining to them that it won't be around until next year because you're still deciding how it will work.

The work on ultrawideband, or UWB, standards has been going on for a while. The technology has been in use for decades. It was considered a military secret early on, because it has applications such as "spotting stealth planes" and "looking through walls." However, the broader demands of commerce are coming into play. Discussions of UWB as a consumer technology don't go back as far, but there are IEEE meetings on the topic going back at least to 1998. The 802.15.3a working group, specifically working on an ultrawideband standard, was authorized in 2002.

The essential conflict is this: two groups of industry leaders have competing ideas of what a final standard for UWB should look like. Last year, the rumor was that a few companies might ship products compliant with one version or another of the spec. In fact, one version of the spec, the DS-CDMA (now called DS-UWB) version pushed by the Motorola®/Freescale™-backed UWB Forum (see Resources), was beginning to see production in early 2004. The competing MBOA (MultiBand OFDM Alliance) version of the spec (see Resources), championed by Intel® and TI™, took a little longer to get ready to roll out, but now both are proclaiming themselves, essentially, ready to ship products. In fact, Freescale, a spin-off from Motorola, is shipping products which may or may not comply with a finalized spec.

Since products for UWB developed by semiconductor companies are generally chips, the time to market of complete standardized products takes a little longer than it might be for already agreed standards; a switch from one spec to the other is probably a new piece of hardware, not just new software to drive it.

The MBOA group has formed a SIG (special interest group) to promote its version of UWB. The IEEE process is still deadlocked. What happened?

The following table will help you keep the UWB players straight:

UWB cheat sheet
NameUWB ForumMBOA
Big BackerMotorola/FreescaleIntel
FCC ApprovalYesNot yet
Chips in productionYesNot yet
Basic technologydirect sequence (DS-UWB)multiband (orthogonal frequency division multiplexing -- OFDM)

What is UWB anyway?

The FCC defines a signal to be a UWB signal if the signal?s bandwidth is at least 20% of its center frequency, or if the center frequency exceeds 2.5 GHz, at least 500 MHz. Either way, the intent is to use a very broad frequency range, so that the signal will not interfere easily with other signals, being so broad and low in transmission power that it looks like background noise to them.

A technology by any other name

Other terms for UWB include impulse, carrier-free, baseband, time domain, nonsinusoidal, orthogonal function, and large-relative-bandwidth radio/radar signals. According to Multispectral, the term ultra wideband was first applied to these systems around 1989 by the US Department of Defense (see Resources).

UWB is supposed to allow you to have many devices communicate wirelessly, with lots of bandwidth and very little interference. Examples include media players, monitors, cameras, and cell phones. Imagine that your flat panel display is somehow able to pick up video broadcast from your DVD player. That, apparently, is the idea. If it takes off, it could conceivably help quite a lot. I have a couple of computers sharing a desk, and it's just a massive nest of tangled wires. UWB, we are told, will fix this.

UWB has many other applications with perhaps greater social importance than detangling cables. For instance, medical imaging, automobile collision-avoidance systems, firefighters and police looking through walls, as well as finding and tracking assets and people. This is a technology which, in at least some applications, could be saving lives.

As it turns out, this is about all anyone will agree on.

UWB signals are essentially scattered all over the spectrum, for example, for communications and measurement applications between 3.1 to 10.6 GHz.. Both parties competing for the IEEE 802.15.3a standard plan to use similar frequency bands. But where the two plans differ is in the type of modulation: MBOA uses frequency hopping (FH-CDMA) and multiple bands at once; the MBOA spec divides a range of frequencies up, allowing devices to use different frequencies in different regions to comply with possibly different regulations, as well as to avoid clashes with other radio devices. The UWB Forum uses DS-CDMA so that everyone shares the whole space (well, actually, they divided it into two bands), using encryption to keep each signal segregated from its neighbors. They have basically different design philosophies, and that's why they're incompatible.

The basic idea is clear, though: a signal spread out so broadly that it just looks like background noise if you aren't the one it's aimed at.


FCC, IEEE, and ITU

Some people are concerned that a standard that would seem entirely acceptable within the United States might not be so welcome elsewhere. Some commentators suggest that the level of background radio noise that is presumed in some versions of this spec would be unacceptable to regulatory bodies outside the US. This kind of problem has been a serious one for previous specifications. 802.11a uses the 5.8GHz band, which the ITU (International Telecommunications Union) had earmarked for a technology called HiperLAN, which was supposed to offer high-performance networking. Similarly, cell phone technologies have often been unportable between the US and Europe, because different bands and technologies have gotten regulatory approval.

What about Europe?

While all the wrangling is going on in the US, the European Union is looking into UWB as well. For now, the European Commission has issued a mandate to investigate this and propose suitable regulatory limits. If things go well, the European regulatory limits will be similar enough to FCC regulatory limits to make globally compatible UWB possible. Unfortunately, ITU and the FCC don't always coordinate particularly. That could leave us with an IEEE standard that can be legally implemented only within the US, with different regulatory limits for any attempt to use UWB technologies in Europe.

There is a similarity between the difficulties of establishing common regulatory limits worldwide, and the kinds of conflict that the factions of the IEEE 802.15.3a committee are currently involved in, and the net result is much the same: a harder time for developers and users alike.

This ties back into the existing standards battle. When the FCC (Federal Communications Commission) essentially told the UWB people to go ahead and start competing in the marketplace, it pretty much gave them a license to establish de facto standards in the US that may be very unpopular in Europe. After a couple of decades during which cell phone technology has been surprisingly unportable, you'd think people would know better! What good does it do you to have a cell phone and a PDA that can exchange data, if they are required by law to be powered off the moment you leave the country? For that matter, this also increases manufacturing costs, and thus consumer costs, decreasing sales.

Andrew S. Tanenbaum once observed that "the nice thing about standards is that there are so many to choose from." Prospective vendors wanting to develop UWB products now have the option of choosing between two competing proposals. Doesn't sound bad enough yet? There are rumors of groups splitting off from those to develop yet another.

Meanwhile, the FCC has endorsed the UWB-enabled chips that Motorola's spinoff, Freescale, is now shipping. They meet US FCC regulations for interference with other devices, and are thus allowed to get out there and start being entrenched. This certainly gives Motorola/Freescale's side of the argument a bit of an advantage.

On the other hand, from a number-of-companies perspective, the MBOA spec clearly has an advantage. Their proposals have tended to do better in votes on the IEEE committee, while still falling short of the 75% majority they'd need to become the official standard. What they don't have, as of this writing, is official FCC approval. There is some discussion about how the noise created by a wideband device should be measured, and the question may still be up in the air. (Breaking news suggests the FCC is authorizing them to go ahead; see Resources.)

Last but not least, some people don't want either of the competing systems because they feel the background noise will be unacceptable. The argument is simple: while any individual UWB device's noise will be practically unnoticable, hundreds or thousands could render whole chunks of spectrum unusable. Since one of UWB's big selling points is overlap with the existing radio spectrum, this could be very disruptive. One analogy is street lights. The effective range of a UWB device is pretty small, not much larger than the area covered by the glow of a typical street light. But, as amateur astronomers have known for years, a few tens of thousands of street lights in a major metropolitan area make it very hard to see the stars. If UWB devices become commonplace, similar concerns apply to the chunks of radio spectrum they operate in.

It's not just radio astronomers. There are worries that UWB signals could impact the channels used by police, fire departments, and other people who place a very high priority on reliable radio communications, and whose signal bands are currently protected. These concerns have yet to be entirely addressed, and the potential downside for vendors is huge. Imagine shipping a million products only to find out that they're interfering with police radio.


Trial by marketplace

One way of resolving a conflict that doesn't seem to be getting any better through the normal standards process is to let the parties fight it out in the marketplace. That is to say, let them ship products, and see which ones eventually win.

This is sort of hard on everybody. It's pretty obvious how end users suffer. They have to gamble on a standard proposal that might lose. That leaves them with an investment in brightly colored plastic with bits of metal stuck in it. (I speak with the authority of someone who still has a Betamax VCR in the attic.) For enterprise users, the risk may be unacceptable, leaving them, not with two (or three) options, but with none.

If one of the proposals wins, the companies that were involved in the losing proposal take a fairly serious hit. All their development time and investment is gone, and they have to design and build new chips. This could take two years or so. But even the winners don't have a free ride. The extra, competitive pressure from the desire to set a standard means they probably haven't been able to make as much money as they otherwise might off of a slow but steady start in a more cohesive marketplace. To cap it off, marketing and promotion expenses are unusually high for both groups.


Spin

Of course, no one can say "hey, you know, the standards process wasn't favoring our business interests over our competitors, so we're doing this other thing instead."

You can, of course, say it about the competitors, and indeed, some of the MBOA members have said that about Motorola, and now about Freescale. On the other hand, the UWB Forum has lots of stirring words about the importance of adopting a "proven" standard. After all, they have working devices and regulatory approval. Both pages have occasional comments on "alternative approaches" or "other designs," by which they seem to mean each other.

Unfortunately, the need to make the products sound enticing has won over detailed technical specifications. You can get real information from the IEEE Web page, but even then, it often falls short of answering the questions potential developers would want answered. Both parties talk positively about Reasonable And Non-Discriminatory (RAND) terms for intellectual property. Both are enthusiastic about time to market and consumer benefits. It makes me feel warm and fuzzy to know how much both of these groups want me to be happy. Both groups (and neutral third parties) generally agree that UWB will be very low power, but neither will tell me exactly how much power one of their chips will consume. One group said that its device could run off a standard AA battery, but didn't say for how long.

Part of the problem, of course, is that there aren't actual final chips in production for one of the proposals, so it's hard to get final technical data.


Is it all necessary?

From the outside, I can't really say if it's all necessary. I don't even know whether, as an end user, I would ever notice the technical differences between these two technologies. It's a very interesting question. It's not just that I'm not sure which of them I'd like better; I'm not sure I could tell the difference at all. End users, and even product developers, probably care more about availability of a standard than which one they get.

There is, of course, not much in the way of detailed comparisons by neutral parties.

Do these different approaches have significant impacts on the power consumption, range, or interference qualities of the device? No one will tell you that. Both groups assure you that their power requirements are quite reasonable, their range is good, and they produce no interference worth talking about. Of course, there's regulatory agencies, which vary from country to country. Both groups, of course, assure us that their version of the standard will have no regulatory problems. Detailed answers to these questions are not being made available right now.

It's pretty clear that both groups are deeply entrenched. What's not so clear is how much of this has to do with the technical merit of their proposals, and how much of it has to do with optimistic investment in the path they thought the technology should take. It may well turn out that there are real-world scenarios where each technology path makes sense, and that the dispute comes from a disagreement over which cases matter the most. In this case, the desirability of a common standard (with interoperability between devices) can be in conflict with the suitability of the competing technologies to specific purposes.

Discussions about the split, and before that, about the predicted split, and before that, about the possible upcoming split, in the UWB standard have been happening for a couple of years. Throughout that time, both sides of the dispute have been, it appears, firmly committed to making their side win. Reading between the lines, it sounds like both parties are committed enough that, even if the competing spec were to become the official 802.15.3a spec, they might well continue to push their own spec, possibly through another standards organization, such as WiMedia (see Resources).

There is some talk of compromises, or of ways for the competing standards to at least coexist, but the uncomfortable reality is that it may take years for the marketplace to pick a leader. Worse, it could end up like something we've seen with cell phones, where different standards win in different countries, leaving us with a massive nest of tangled wireless. One compromise proposal involved every device simply having both physical network access layers, or PHY chips, so they could negotiate. This is not exactly appealing, whether to vendors or consumers.

Clashes like this have been resolved in various ways in the past. In software standards, "we'll just provide both" is a comparatively viable solution. For a hardware standard, it's not so appealing. There is still some hope that one side or the other will blink first, resulting in a single, unified standard, or possibly that both sides will make compromises and yield a single, unified design everyone can live with. An organization like WiMedia, the not-for-profit that has been trying to keep all the 802.11 people interoperating, might be able to help. Unfortunately, right now, the biggest players are very much entrenched in their positions.

When an issue with a standard simply can't be resolved, sometimes that means it's time to step back and look at the assumptions that went into the standard. In this case, the major players felt -- and continue to feel -- the weight of time-to-market and economic pressures. In the meantime, developers can go ahead and start off with design kits based on either of the competing proposals. If you don't need to interact with anyone else's technology, either is fine for now. If you need interoperability, it's probably best to wait for an official standard.

Many thanks to Dr. Walter Hirt of IBM Zurich Research Laboratory, for reviewing and commenting on a draft of this article.

Resources

Comments

developerWorks: Sign in

Required fields are indicated with an asterisk (*).


Need an IBM ID?
Forgot your IBM ID?


Forgot your password?
Change your password

By clicking Submit, you agree to the developerWorks terms of use.

 


The first time you sign into developerWorks, a profile is created for you. Information in your profile (your name, country/region, and company name) is displayed to the public and will accompany any content you post, unless you opt to hide your company name. You may update your IBM account at any time.

All information submitted is secure.

Choose your display name



The first time you sign in to developerWorks, a profile is created for you, so you need to choose a display name. Your display name accompanies the content you post on developerWorks.

Please choose a display name between 3-31 characters. Your display name must be unique in the developerWorks community and should not be your email address for privacy reasons.

Required fields are indicated with an asterisk (*).

(Must be between 3 – 31 characters.)

By clicking Submit, you agree to the developerWorks terms of use.

 


All information submitted is secure.

Dig deeper into developerWorks


static.content.url=http://www.ibm.com/developerworks/js/artrating/
SITE_ID=1
Zone=Multicore acceleration
ArticleID=33527
ArticleTitle=Standards and specs: A house divided
publish-date=01252005