The 802.11n story: The book I wish I'd written

by C.J. Mathias

December 12, 2005 —

As regular readers of this column know, I love to write. I grind out thousands of words every week, and, absent the occasional case of writer's block, I love to do it. I've been approached by a number of publishers about doing a book, and, while I'm indeed flattered to receive such offers, I can't commit to an effort of that scale. I read a lot of books, mostly on wireless technology, and most of them suffer from Too Many Words syndrome. A really good book, in my humble opinion, exhibits an economy of words, while still using a lot of them to convey the good stuff. Too many books are redundant both internally and externally; I don't have time for that.

But I couldn't do a book, not because I'd commit this sin, but rather because I just can't carve out the time to do it. I am always incredibly busy, and I really like it that way, despite the fact that it imposes certain limitations on other longer-term activities. But one book I wish I'd written would be called The History of IEEE the 802.11n Standard. This tale has it all - technology, intrigue, politics, business, and strategy. If you read Tracy Kidder's classic The Soul of a New Machine, then you know what I'm getting at. If you didn't, this was the story of all that went into building a new computer, from scratch, back when such things could not be done by mere mortals. The twists and turns in the story are remarkable, even for non-techies.

Such it is with 802.11n. .11n is a task group (TG) charged with producing a standard for the next-generation physical layer (PHY). The key goals are performance above 100 Mbps, and backwards compatibility with the current PHYs. While developing such technology, especially in the form of a standard, is never easy, the path to .11n has been borderline tortuous. And the reason for this has little to do with the admittedly-difficult technology, and things may in fact get worse. Let me explain.

The most important radio technology today is MIMO. MIMO, which stands for multiple-input, multiple-output, adds improved range, reliability, and throughput to wireless through the use of spatial techniques. This is like adding a third dimension to the frequency (bandwidth or spectrum) and time already used, and more dimensions mean more room. That's it in a nutshell.

As you might guess, though, MIMO is a very, very difficult technology to implement. It has been used over the years in a variety of systems, but usually government and military projects where cost is of lesser importance. While I started looking at MIMO in 1995, even I was stunned when the first MIMO-based WLAN products came along in 2003. This was Belkin's "Pre-n" product line. Tests quickly showed that it was at least twice as fast as ordinary .11g products, and, while expensive, it clearly pointed the way. MIMO was then and remains today the clear and obvious future of WLANs.

The .11n standards effort also began in 2003. Note, however, that standards are not products. Rather, they are rigorously-defined specifications for the interfaces between functional elements. This allows, for example, a WLAN AP from one company to interoperate with a client adapter from another. And while manufacturers can (and usually do) produce products with extensions that can result in incompatibilities, most industry players realize that we're better off with standards than without. Beyond interoperability, standards bring us lower costs and, most importantly, a warm, fuzzy feeling that our investments will in fact work in production applications and be financially protected, at least for a while.

So, what's the problem with standards? Well, the biggest one is that the people working on these themselves work for companies that are competing with one another. And while they labor to produce what amounts to a level playing field, that's bad news for the marketing types, who live for the sustainable competitive advantages that come from clear product differentiation. Commoditization is the enemy; it drives prices lower. And there's a real danger that one of the players working on a standard might have a fundamental advantage, especially with respect to time-to-market, as a result of decisions made by the standards body.

And that appears to be what's happened with 802.11n. In a nutshell: Airgo Networks jumped to an early lead in MIMO simply by doing the very, very difficult: building the first commercial MIMO WLAN chipset. Airgo has continued to push the possible but very, very difficult, and is now into its third-generation chipset, which offers 240 Mbps of raw throughput and well over 100 Mbps sustained. Others, including WLAN-industry leader Atheros Communications, are beginning to catch up, producing chips that are winning in some benchmarks. I expect every WLAN chip and systems supplier to be in the MIMO/WLAN game shortly - that's clearly where the future is. That's what 802.11n is really all about. MIMO is the only reasonable path to higher performance, short of more bandwidth (see my earlier column, "What is Performance?", for more on using additional bandwidth for higher throughput) - and MIMO could be applied to very-wideband applications as well.

But I digress. It's been a thorn in the side of WLAN chipset firms like Atheros, Broadcom, Marvell, and especially Intel to see Airgo jumping to the lead in what is clearly the most important direction for the entire WLAN industry. So, these companies have spearheaded the Enhanced Wireless Consortium (EWC) to define their own direction and perhaps slow Airgo just a little. Airgo and its friends and colleagues continue to pursue what is known as the Joint Proposal, or JP, an effort sanctioned by 802.11n to produce a candidate for a draft standard, and have otherwise not moved to directly counter EWC.

Now, there's nothing wrong with EWC bringing forth an alternative to the JP. Such efforts should be encouraged, within the context of a standards effort; it's always best to get as much technology on the table as possible before casting a standard in concrete, er, silicon. But is the fundamental purpose of EWC to simply gain back momentum lost to Airgo, and will the backers of EWC produce products based on it even if EWC is not the 802.11n standard? The latter is a real possibility, because building chips - especially those with the complexity of MIMO - requires long lead times. Even Intel can't afford to start building chips and then abandon the effort when an alternative, even in the form of a standard, comes along. While such redirection does in fact happen from time to time, it's expensive. And, by the same token, it's very difficult for a major chip supplier to wait for a standard, because, again, lead times are long and momentum can't be lost, especially when competitors are already building significant marketshare and mindshare.

And that's where we are, right now. I would even go so far as to say that there is a possibility, although with a probability of less than 50 percent, that there won't be an 802.11n standard, at least not any time soon. We might see a bunch of incompatible MIMO products from a broad range of vendors, just as we do today in the residential market. This is acceptable in the home because all of these products fall back to the 802.11g standard, and with better performance than vanilla .11g implementations, and because multi-AP, managed installations are very rare in residential neighborhoods. But enterprises aren't going to be crazy about the lack of a standard; the existence of a standard is usually the catalyst for the "hockey stick" uptick in demand for new or otherwise exotic technologies - sales muddle along for promising new approaches until standards really bring in the customers.

I have always believed, to quote someone else, that "there is no limit to what can be accomplished provided one is unconcerned about who gets the credit." Unfortunately, there's not much profit potential is this line of thinking. Competition is, in many ways, a good thing, improving quality, lowering prices, encouraging innovation, and being a major factor in the amazing variety of goods and services available in market economies. The downside, however, is that it can often have exactly the opposite effect when powerful players are threatened. It's still too early to tell how all of the above will turn out, and I remain optimistic that not only will .11n happen, but there'll be a slew of standard MIMO products on the store shelves by next Christmas. Of course, I could be wrong - and I still think, regardless, that this story would have made a really great book.

Farpoint Group