Features

Broadcom achieves diplomatic fudge over "compatibility" testing

Just when you thought the WiFi Alliance was going to play tough with Broadcom over its launch of a pre-11g wireless LAN standard, the California-based chip designer has pulled off a publicity coup by getting "WiFi accreditation for its 54g chip set." But is this what it seems?

It's true! But does it mean what you imagine it means? Well, that depends on what you imagine.

Try this: "Consumers are rapidly adopting high-speed wireless LAN products based on our 54g(TM) chipsets, and WiFi certification for our reference designs will provide them with added assurance that 54g(TM) equipment will interoperate with their existing 802.11b devices," said Jeff Abramowitz, Senior Director of Marketing for Broadcom's wireless LAN products. (That's his press release, by the way). He goes on:

"Our latest OneDriver software release also provides a significant industry confidence boost in 802.11g draft specification products. We have demonstrated interoperability to 54 Mbps with all competing chips on the market, and this release confirms expectations that 802.11g draft specification changes can be accommodated with a simple software upgrade."

What you may imagine, is that this is a resolution of the dispute between some Alliance members, and Broadcom. You might recall reading stories saying that some people were really annoyed with Broadcom for launching chips based on the 802.11g spec before the spec was finalised. You might also recall stories saying that some engineers - whether they are speaking on behalf of the Alliance or not - think that Broadcom has mis-read that preliminary spec.

And you might imagine that all that unpleasantness is over.

What has actually happened is that someone has lifted the carpet, and is operating the broom. All that we know for sure is simply that the 54g chips are certified compatible with 802.11b chips. They've been certified, therefore, as working as ordinary 11b - that is, 11 megabit per second - wireless LAN systems.

What we need to know, is what Abramowitz means by "We have demonstrated interoperability to 54 Mbps with all competing chips on the market."

Does he mean that this was part of the WiFi Alliance testing procedure? Then he should say so. It's highly unlikely, since only a week ago, the Alliance said that "WiFi certification testing of IEEE 802.11g products will begin after the IEEE has approved the final standard."

That isn't expected till Summer at the earliest, with pessimists saying "could be as late as October."

Does he mean that there have been privately-arranged interoperability tests between his company and Intersil, the only two chip designers currently shipping pre-11g products? If so, what exactly was tested?

Most important: will these "11g" products actually run at 54 megabits even when there are 11b wirelesses operating in the area? And when he says: "this release confirms expectations that 802.11g draft specification changes can be accommodated with a simple software upgrade," does this guarantee that this is one of the changes?

His announcement came out on Wednesday last week, and since then, a conspicuous lack of information has emerged.

And what else has happened since then is that a lot of busy journalists have simply printed the Broadcom press release.

Compare the following:

Here's the original -

PR NewsWire.

And here are the "news reports"

http://www.80211-planet.com/news/article.php/1676951

http://biz.yahoo.com/prnews/030226/law062_1.html

http://www.unstrung.com/document.asp?doc_id=28957

http://www.commsdesign.com/pressreleases/prnewswire/59678

See if you can see the differences:

"IRVINE, Calif., Feb. 26 /PRNewswire-FirstCall/ -- Broadcom Corporation (Nasdaq: BRCM - News), the leading provider of integrated circuits enabling broadband communications, today announced that reference designs and customer products based on the 54g(TM) wireless LAN chipset have achieved WiFi® certification for 802.11b interoperability. The WiFi certification demonstrates that wireless LAN products based on 54g(TM) technology can achieve 100% compatibility with existing 802.11b products while also providing almost five times the bandwidth when operating in 54g(TM)-only environments. 54g(TM) is Broadcom's maximum performance implementation of the IEEE 802.11g draft specification."

"IRVINE, Calif. -- Broadcom Corporation (Nasdaq: BRCM - message board), the leading provider of integrated circuits enabling broadband communications, today announced that reference designs and customer products based on the 54g wireless LAN chipset have achieved WiFi certification for 802.11b interoperability. The WiFi certification demonstrates that wireless LAN products based on 54g technology can achieve 100% compatibility with existing 802.11b products while also providing almost five times the bandwidth when operating in 54g-only environments. 54g is Broadcom's maximum performance implementation of the IEEE 802.11g draft specification."

"/PRNewswire-FirstCall/ -- Broadcom Corporation (Nasdaq: BRCM), the leading provider of integrated circuits enabling broadband communications, today announced that reference designs and customer products based on the 54g(TM) wireless LAN chipset have achieved WiFi(R) certification for 802.11b interoperability. The WiFi certification demonstrates that wireless LAN products based on 54g(TM) technology can achieve 100% compatibility with existing 802.11b products while also providing almost five times the bandwidth when operating in 54g(TM)-only environments. 54g(TM) is Broadcom's maximum performance implementation of the IEEE 802.11g draft specification."

and so on. If there are differences, they elude quick detection; the only difference seems to be whether the cut-and-paste job included the ascii character for "(TM)" or just printed the brackets and letters.

It's hard to see how the WiFi Alliance could have prevented this; if someone launches a product which passes all its tests, it has to approve it.

On the other hand, surely, it must be obvious that the press release doesn't make anything any clearer than it was before? We know - what we have always known! - that pre-11g products work with 11b products. We also know - again, no news! - that Broadcom believes that any incompatibilities can be ironed out in the future by a software upgrade.

And surely it must be equally obvious that what Broadcom has pulled off, here, is a successful fudge of the issues, which might well confuse many customers?

Some of us believe that the problems, such as they are, are pretty trivial. Some of us reckon that if you're installing WLAN equipment at home, you are pretty safe installing a 54g product, even if it doesn't run at full speed with 11b devices in the area. We might go so far as to say: "Come on, it's no big deal; even if you buy a preliminary spec wireless device now, and you have to buy a new one in a year or so, you're only looking at a hundred dollars worth of spend for the typical household!"

I might myself go further and say: "Faced with a choice of a Linksys 11b and an 11g(pre) 54g router, go for the 54g - it may not be a standard product yet, but it works better, even for users of 11b devices." In a forthcoming review, we will in fact say something like this.

On the other hand, if it's truly a trivial problem, surely the people involved could afford to be open and candid about it?

Does Broadcom actuallly understand why what it has done here doesn't seem open and candid?

Let's hope that some of these questions turn out to be answerable.