This article originally appeared in TidBITS on 2006-12-11 at 12:00 p.m.
The permanent URL for this article is:
Include images: Off

Do 802.11n Chips in Macs Signal New Wireless?

by Glenn Fleishman

Shortly after the new MacBook Pro with the Intel Core 2 Duo processor [1] was released, a forum member took a close look [2] and found, among other revised specifications, an 802.11n chipset from Atheros. Earlier, another forum member wrote about [3] finding a Broadcom 802.11n adapter in the Core 2 iMacs.

802.11n is a wireless-networking standard still in formation at the IEEE, an engineering standards body. The 802.11n standard will supersede, but will be backwards compatible with, 802.11b (which Apple calls AirPort) and 802.11g (AirPort Extreme), both of which are part of Wi-Fi. (Wi-Fi itself is a mark that indicates a piece of hardware has been tested for interoperability and for meeting specific tests.)

Although 802.11b runs at 11 Mbps and the 802.11g specification runs at 54 Mbps, those are raw network speeds, which include all the networking overhead that enables chunks of data to be wrapped up into packets and sent over the air, including the bits used when radios interact. 802.11b really produces about 5 Mbps of throughput, and 802.11g, without a host of proprietary extensions that some manufacturers include, hits about 25 Mbps of real throughput.

By contrast, 802.11n will be available in raw speeds from 150 Mbps to 600 Mbps; real world throughput will start at 100 Mbps and is expected to reach 300 to 450 Mbps in the most expensive devices with all the optional bells and whistles. 802.11n also requires MIMO (multiple-in, multiple-out) antenna arrays that have been in products for a couple of years. MIMO antennas dramatically increase network range and improve throughput at shorter ranges.

The problem with Apple's apparent inclusion of 802.11n at this point in time is that no standard exists. Several chipmakers decided in early 2006 to release silicon based on the first working draft, called Draft 1.0, from 802.11's Task Group N, the group deliberating on the standard. Draft 1.0 appeared after more than a year of horse trading and even the near dissolution of the task group, which could have thrown the wireless networking world into slight disarray. But Draft 1.0 is just what is sounds like: a draft.

This Draft 1.0 silicon may be rather different from the final standard. And there's no guarantee that hardware upgrades for any so-called "Draft N" equipment sold this year will work with the final, approved standard - or even with future drafts! (Asus is the only company to offer a guarantee of replacement hardware, but not until 2008, when a final standard is expected.) That is, Draft N chips released now might work with each other (that's not guaranteed, and is considered one of the current big drawbacks), but they might not interact with future, true 802.11n devices at speeds faster than the fallback of 802.11g.

Right now, Draft 2.0, incorporating hundreds of technical comments on the first draft, is expected in January 2007 with approval in March 2007. This draft would then serve as the basis of a plan by the Wi-Fi Alliance [4], the group that tests and certifies Wi-Fi-marked equipment, to ensure interoperability within a few months of that point in time - almost certainly before June 2007. This interim certification program would give some market stability while the standard moves toward expected full completion by early 2008.

Upping the hype was a recent announcement by Qualcomm on the same day it acquired MIMO pioneer Airgo. Airgo's MIMO chips incorporated many principles of 802.11n and helped set the direction of 802.11n. Qualcomm said that Airgo was announcing the "availability" [5] of Draft 2.0-compliant chips. Through interviews later, the company clarified [6] that, first, "available" meant "in real quantities for producing devices after March 2007;" and, second, that "Draft 2.0 compliant" meant that at this stage in the game, all the parameters that might be in Draft 2.0 are known, and Qualcomm's new division claimed to have all those parameters in their chips. This is a more reasonable statement, because even with many technical comments left to be resolved by the task group, it's a manageable pile with probable outcomes.

More interesting, however, is Qualcomm's claim that they also support Draft 1.0, which could mean that Qualcomm would have Draft N devices that would prevent even gear made with other companies' chips from a kind of obsolescence when real Draft 2.0 devices ship.

For Apple to include Draft N silicon now is therefore baffling, with the only chipmaker pretending to have something that resembles Draft 2.0 not available for manufacturing products until the second quarter of 2007. I could see them adding MIMO as an overlay on 802.11g, which is rather typical in many products now, and waiting until at least early 2007 for a certified draft version of 802.11n.

If Apple chooses to enable the Draft N features when the iTV media adapter ships in early 2007, there's no guarantee that future Draft N chips would have full backward compatibility with what they ship. While many people compare the early release of Draft N devices with 802.11g, which appeared in equipment from Apple, Linksys, and others several months before its ratification, 802.11g was past Draft 5.0 when the first chips shipped, and had only minor changes after that point. And even those changes prevented interoperability of 802.11g equipment from different firms using the same wireless chips initially; Apple released something on the order of six firmware upgrades between AirPort Extreme's shipping date and 802.11g's final approval at the IEEE.

Apple often pushes the envelope, but if it proves true that they're this far out ahead, they might be tearing that envelope.