The Ratification of 802.11n: Planning for Full-Speed WLANsThe Ratification of 802.11n: Planning for Full-Speed WLANs

A number of planning steps are essential for getting the full impact of an 802.11n implementation.

Earlier this month the IEEE finally ratified the 802.11n standard providing a better and faster wireless LAN (WLAN) radio link. Despite the marvelous technology it employs, the announcement was met with a collective yawn, a wonderfully apt response. The lackadaisical reception is based largely on the fact that the Wi-Fi Alliance has already certified over 200 devices as complaint with the Draft 2.0 version of the standard, user deployments are well underway, so the formal ratification was a rather unremarkable denouement. But while the ratification isn't unexpected, comprehensive deployment of 802.11n systems will require some new and enhanced effort on the part of enterprises.

Wi-Fi Alliance Succeeds Again
Normally enterprise users would be reluctant to risk going off half-cocked with a "pre-standard," but the Wi-Fi Alliance had guaranteed that products built and certified to the Draft 2.0 standard would be backwards compatible with the eventual standard when it was ratified. Based on that guarantee and the Alliance's track record of delivering on its promises, many organizations, particularly those in WLAN-intensive environments like education, have already begun their 802.11n deployments.

The great thing about Wi-Fi and the reason for its success is that the process works just the way these technology things are supposed to work. That is due in no small part to the cooperative roles of the IEEE and the Wi-Fi Alliance. The IEEE develops the official standards, which process is governed by a stringent set of rules that ensure "due process, openness, consensus, and balance" with a right of appeal at each step along the way. That process takes time, and results in standards that may include too many options to ensure they will interoperate reliably. Tha't’s where the Wi-Fi Alliance comes in.

For each major development in the 802.11 standards, technical committees in the Wi-Fi Alliance define a certification program to ensure compatibility. Those functions are typical of all vendor alliances or forums, but the Wi-Fi Alliance has taken it a step further. When there is a pressing need in the market, the Alliance will "jump the gun" on the standards process, which is essentially what they did in developing a certification for the Draft 2.0 802.11n standard. However, their track record of ensuring compatibility while fostering innovation and maintaining backwards compatibility has made "Wi-Fi Certified" a guarantee that businesses and consumers can count on.

User Planning Steps
The technology in 802.11n is a major step forward and it addresses the Three-R’s in radio: rate, range, and reliability. Without a doubt, 802.11n will eventually supplant the earlier 802.11b and g interfaces, and it will likely put a cap on the less popular 5-GHz 802.11a interface. While 802.11n can operate in either the 2.4-GHz or 5-GHz bands, most users have yet to deploy 802.11a so the 5 GHz band is virgin territory. With the number of channels available in the 5-GHz band, it will accommodate the 40-MHz double-wide channel option in 802.11n; in a 4x4 MIMO implementation, those 40-MHz channels will support data rates up to 600 Mbps.

Many of my clients have already started to move to 802.11n, and always in that 5-GHz band. While 802.11n is backwards compatible with the earlier radio links, operating in that Mixed-Mode results in a serious performance degradation for the n-devices. That degradation is due in part to the fact that lower speed users will be sharing the channel with higher speed users, but it also negates many of the MAC-layer protocol enhancements that were included in 802.11n.

There are a number of other planning steps that are essential for getting the full impact of an 802.11n implementation.

First, you will likely need a new site survey to determine optimum access point placement. Besides the different loss characteristics of a 5-GHz versus a 2.4-GHz radio signal, the MIMO transmission used in 802.11n results in a coverage pattern that is far different from a traditional access point. Test tools will also need to be updated, and we have found that basic signal strength and signal-to-noise ratio (SNR) measurements do not provide an accurate gauge of 802.11n transmission rates. Your wireless intrusion detection/intrusion prevention system (WIDS/WIPS) will also need an upgrade or it will not recognize rogue 802.11n APs.

As we can anticipate data rates in excess of 100 Mbps, a single Fast Ethernet interface to the AP will no longer be sufficient. In the past, WLAN transmission rates were so slow relative to wired connections that we could almost treat them as "background noise." With the dramatic increase in wireless transmission rates we will now have to focus more attention on the wired infrastructure to prevent bottlenecks.

A potentially more costly issue is the fact that many 802.11n access points cannot operate on the 12.95 Watts delivered over an 802.11af Power over Ethernet interface. What we are finding is that vendor strategies to accommodate traditional PoE may result in lower transmission rates--buyer beware.

Finally, while much has been made of 802.11n's impact in terms of WLAN voice devices, we have yet to see any 802.11n-capable WLAN handsets. Given the power requirements of the 802.11n chipset, voice devices will likely use a 1x1 implementation, which provides a maximum transmission rate of 72 Mbps (on a 20 MHz channel). Interestingly, Apple has reportedly used an 802.11b chip (i.e. the Broadcom BCM4329) in their new iPod Touch, but without the firmware that would allow it to run 802.11n. Given the battery problems Apple already has with their iPhones, I'm not optimistic about an n-capable iPhone in the near future.

What's Next?
It will still be a couple of years before we have devices that deliver the full impact of 802.11n. Most chipsets today are 2x2 or 3x3, so you are not getting the maximum data rates available with a full 4x4 implementation. While 4x4 chipsets have been announced, it appears that most users (and not a few wireless "professionals") are still not clear on the various implementations for 802.11n or their impact on performance.

In the longer term, there are two higher-speed Wi-Fi interfaces in the works, and those should push the raw data rate over 1 Gbps. They are 802.11ac, which will operate in bands below 6 GHz; and 802.11ad, that will operate in the 60 GHz band. The signal losses at 60 GHz are such that the 802.11ad interface will likely be limited to in-room communications, where the major applications might be in consumer entertainment devices. Remember, Wi-Fi isn't strictly for business.

On the whole, the WLAN switching business has gone from being a technological marvel to a rather boring infrastructure element, but there are some interesting developments brought about by 802.11n. WLAN switching vendors are finding that scheduling algorithms that give precedence to higher-volume users can have a significant impact on overall network throughput with little performance impact for lighter volume users. So a subtle design feature may have a major impact on performance, at least in some implementations.

As we look to the future, there are likely two major wireless interfaces that will be required in a full-function mobile device: 802.11n and LTE. Some might add Bluetooth to that list, but with the capacity available in 802.11n, it could conceivably support a headset connection as well as a network connection; we’ll have to see how that plays out based on cost and power requirements.

While the cellular carriers would like us to think that LTE will meet all of our needs, given their available radio spectrum and the limitation imposed by Shannon’s Law, that is not likely to be the case. Cellular devices are increasingly being equipped with Wi-Fi capability, so why settle for a piddling 6-zeros for a transmission rate when you can get 7- or 8-zeros!

Like all major technical innovations, 802.11n will raise questions for users. So while 802.11n will provide the brawn, it will still require the brains to achieve its full potential.