[Please note: This editorial was written in 2006. In 2007, I changed my advice when Draft N was finalized into version 2.0, with changes incorporated into shipping hardware. Then the Wi-Fi Alliance added a pre-ratification certification process. I fully recommend Draft N gear now. –Glenn Fleishman, 2009.]
I am going to make myself extremely unpopular this morning by suggesting that no one buy so-called Draft N Wi-Fi gear that is pouring into the market: Buffalo and D-Link say they've been shipping wireless equipment based on the current draft of 802.11n, the higher-speed cousin to existing Wi-Fi, for at least two weeks. Linksys entered the fray today with a new router and PC Card. Other manufacturers who have not yet announced will follow in May and June. And it's a bad idea.
These companies are shipping Draft N devices for the bragging rights to be first out of the gate and to try to brand their Draft N products' identities on consumers' minds. There's no good technical reason to release these products this early. Lest someone point to 802.11g's early release before ratification by the IEEE, it's important to note that 802.11g was several drafts further along when Broadcom pushed out its first chips.
I came around to the idea of early multiple-in, multiple-out (MIMO) equipment more than a year ago because MIMO gear, once it started to drop in price, offered the significant advantage of dramatically increased coverage and range even while using older Wi-Fi adapters. For specific purposes in which you needed more throughput for a wireless link, purchasing all new MIMO gear also made sense as a limited investment with no path to ensure that the higher speed of that equipment would work with any future standard. So far, so good.
But Draft N gear makes no sense for consumers because the actual standard is close enough at hand that buying pre-standard equipment has little benefit and a possibility for disruption given manufacturers' current guarantees and claims. Here's the lay of the land.
No guarantee for upgrade. There's still a small but significant risk that equipment shipped today will contain chips and firmware that cannot be upgraded to provide full compatibility with the final standard. Buffalo was brave enough to say in an interview two weeks ago that they are not guaranteeing that their Draft N equipment will be upgradable, although they believe it almost certainly will. Other manufacturers are not so brave, but they are also not offering a replacement guarantee.
With no hardware replacement guarantee from a manufacturer, why buy today? If manufacturers are willing to step forward and provide explicit offers with their products that they will have the same compliance with the gear they shift today as with the gear they ship in six months, up to and including replacing the hardware you purchase today, then that argument goes away.
No agreed-on, tested plan to avoid interfering with older networks. A more significant concern is that chipmakers and the IEEE task group working on 802.11n haven't agreed on a single approach to mitigate interference with existing legacy Wi-Fi networks, nor are the individual methods well tested ina real-world environment or with each other.
Over a month ago, I wrote about Airgo's self-motivated concerns in pressing this point, but I'm now convinced that they're generally correct about the issue, although they certainly overstated some of the particulars. Without agreed-upon methods of avoiding interference, many different chipmakers and equipment manufacturers devices will be contending with an inevitable variation in success.
The idea that consumers will apply several firmware upgrades to solve problems that emerge from this contention is a little ridiculous. Although early adopters will likely buy this Draft N gear, there's a lot of evidence to suggest that a device, once deployed, is rarely patched unless someone is advised to do so by technical support or they're a fairly technical user to begin with.
Testing of the first Buffalo and NetGear devices by Craig Mathias of the Farpoint Group, showed poor interoperability, security model problems, and performance. Mathias also recommends avoiding this wave of draft products. (News.com picked up this story and pointed to eWeek's evaluation showing poor range in a test device and legacy network problems, too, but praising the potential. They also noted that Linksys will release the same product with Atheros, Broadcom, and Marvell chipsets depending on availability. Hello, tech support nightmare.)
Gigabit Ethernet is necessary for the full benefit. If you're planning to buy 802.11n equipment for coverage area, this isn't an issue, but if you're not buying it just for range, you need a gateway that has Gigabit Ethernet (1000 Mbps) built in. Only NetGear is offering models that support this speed in these early stages. GigE has become less expensive, but it will take market interest and some months before it's available for a reasonable price.
Prices will plummet within months. If you buy Draft N gear today, you're paying too much. Five major Wi-Fi chipmakers will contend for business in Draft N. Airgo said they're waiting a bit longer to release chips and Intel's plans are unknown, but Atheros, Broadcom, and Marvell are already producing silicon. With five chipmakers and dozens of manufacturers, competition will be intense and prices will fall fast.
Wait. Unless you have a very particular, very short-term need for the highest possible speed using a wireless network, there is little reason to suffer through the shakedown that will take place over the next several months. At the very least, wait until after the May IEEE meeting, at which point it may become crystal clear whether any minor changes to the draft of 802.11n will require major changes in silicon. Even if there are no chip changes, you might as well wait for the inevitable firmware upgrades that will chase the standards group meetings and take weeks to appear following those meetings.
I won't recommend buying Draft N gear until almost certainly September or later--later if the prices haven't fallen substantially by September.