MaximumPC 2006 09

In our last 802.11n white paper (November

2005), we concluded that none of the then-

current MIMO implementations would have

any signifi cant resemblance to the fi nal

802.11n spec. Our prediction remains on

target: The IEEE fl at-out rejected the 802.11n

Draft 1.0 proposal on which the latest batch

of MIMO routers, dubbed “Draft N,” are

based. That doesn’t mean you should give up

on 802.11n; just don’t put a lot of faith in any

“real soon now” promises.

JOCKEYING FOR POSITION

Last year, two camps were competing to

defi ne 802.11n: WWiSE (whose members

included Airgo, Broadcom, and Texas

Instruments) and TGNSync (led by Intel,

Atheros and Marvell). These groups eventu-

ally merged into one called the Joint Proposal

(JP) Team; alas, the unity was short-lived

and Intel, Broadcom, Atheros, and Marvell

splintered off to form the Enhanced Wireless

Consortium (EWC).

In March of this year, the IEEE ordained

the EWC’s proposal as 802.11n, Draft 1.0. But

when the entire 802.11n Task Group voted

by letter ballot, the proposal not only failed to

win the 75 percent super majority required for

ratifi cation, it couldn’t even muster a simple

majority vote in favor. In the meantime, 12,000

comments from IEEE members poured in—10

times more than were attached to 802.11g’s

fi rst draft proposal—and a preponderance

cited defi ciencies or outright confl icts in the

proposed standard.

The backers of Draft 1.0, many of whom

rushed “Draft-N” products to market hot

on the heels of its approval, would have

you believe that all this commotion is to be

expected in such a hotly contested stan-

dard. A look at the actual comments paints a

slightly different picture.

WHATSAMATTA U?

Theoretically, only three of the 11 20MHz

channels in the 2.4GHz frequency band, in

which the 802.11b and 802.11g gear oper-

ates, don’t overlap. These are channels

1, 6, and 11. 802.11n Draft 1.0 specifi es

a 20MHz channel in this same frequency

band, but it also allows for an optional

40MHz channel. This wider signal has the

potential to overlap every channel, disrupt-

ing other wireless-networking communica-

tions in the process. It’s a disaster in the

making for existing Wi-Fi networks.

Thus, many of the comments accom-

panying the letter ballots argued that if the

standard was to permit the use of a 40MHz

channel, it should also dictate a means

of sensing legacy networks operating in

the vicinity and then falling back to a nar-

rower channel. This notion is known as Clear

Channel Assessment (CCA), and it won’t

necessarily hobble an 802.11n network. A

40MHz channel doesn’t need to be entirely

vacant in order to be considered clear; it just

can’t be occupied during the particular mil-

lisecond at which a packet is being sent over

it. Think of speech: It might sound as though

you’re mouthing words nonstop, but there

are countless tiny moments of silence as you

breathe and form syllables.

As simple as it sounds, resolving that one

added feature would require each 802.11n-

compliant device (every router, every adapter,

and every streaming box) to carry additional

hardware to perform CCA and to modulate

the bandwidth—even if the fi nal standard

permits a 40MHz channel but doesn’t require

it (40MHz channels aren’t allowed in Japan

and in some parts of Europe). Making 40MHz

channels and CCA part of the offi cial 802.11n

standard will also render any Pre-N/Draft-N

gear instantly obsolete.

Many of the other comments criticizing

the Draft 1.0 proposal cited concerns about

its capacity—or lack thereof—to stream

audio and video, and its apparent lack of

power-handling controls for handheld devic-

es. Taken as a whole, it becomes apparent

that the IEEE membership views 802.11n as a

standard suitable for all hardware—from PCs

to handhelds and even to appliances—which

will enable wireless LANs to do much more

than anything they’ve done in the past.

NOW WHAT?

At this point, the only difference between all

the existing pre-802.11n hardware is semantic:

Anything that hits the market before the IEEE

approves the fi nal 802.11n standard should

be assumed incompatible with the gear based

on the eventual standard. What we do know

is that the existing pre-802.11n gear had

September marks the third

year of the 802.11n standard’s

gestation, during which time

we’ve seen three entire genera-

tions of ‘Draft-N’ and ‘Pre-N’

hardware. Here’s a look at

where this wireless protocol is

at, and where it’s headed.

BY BILL O’BRIEN

r&dBREAKING DOWN TECH —PRESENT AND FUTURE

68 MA XIMUMPC SEPTEMBER 2006

White Paper: An 802. 1 1n Status Report

Whatever the fi nal 802.11n standard ends up being, spatial multiplexing MIMO (multiple input/multiple output) is likely to be one of its

key features. With this technology, data coming into the wireless access point is split into streams and sent to digital signal processors.

The DSPs multiplex the data and pass it along to a set of radio transmitters. Radio receivers in the wireless adapter send the streams

to a digital signal processor, which de-multiplexes them. The bits are then recombined and delivered to the receiving PC.

HOW IT WORKS Spatial Multiplexing MIMO

WIRELESS ACCESS POINT

DSP

DSP

DSP

Bit

Splitter

Radio

Radio

Radio

Data

WIRELESS ADAPTER

DSP

Radio

Radio

Radio

Bit

Merger Data