Wi-Fi eyes global domination

If your reaction to last week's ratification of the 802.11n standard for Wi-Fi at 100Mbps and beyond was a loud yawn or a "what took them so long," you'd certainly have some justification.

Test of 802.11n access pointsIEEE approves 802.11n Wi-Fi standard

After all, it took the IEEE seven grueling years to finalize the standard. And so-called "pre-standard" products have been shipping since 2007. In fact, wireless vendors report that the majority of their current shipments are pre-standard 802.11n gear.

But let's not gloss over what a stunning accomplishment this really is. Or what it will mean down the road in terms of wireless technology becoming the dominant network technology in enterprise shops.

First, a tip of the hat to the hundreds of participants from across the wireless industry and beyond who hammered away for a good five years on the nitty-gritty details of the standard. (Read a timeline of 802.11n's milestones.)

A lot of the time it wasn't pretty. But the result is a total re-write of the original 1997 wireless standard that puts in place the underlying technologies that will carry the WLAN industry and its growing base of hundreds of millions of users for the foreseeable future.

While there are other radio technologies under development, nothing will replace 802.11n anytime soon. But, in time 802.11n will replace 802.11b and g, while still providing backwards compatibility with these earlier standards.

802.11n delivers improved throughput, range, and reliability across a broad range of configuration possibilities, and thus cost and price points.

Implementations can be basic -- one radio on the transmitting side and one on the receiving side offering up to 150Mbps, as opposed to the 54Mbps of 802.11g. Or more complex -- the commonly deployed 2x2 and 2x3 configurations can deliver up to 300Mbps, and a 4x4 implementation offers up to 600Mbps (all of these being peak numbers).

Prices of 802.11n products have fallen dramatically, as is always to be expected where VLSI chip manufacturing is concerned, to the point where 802.11n products today cost about the same as chips based on 802.11g just two years ago.

Built-in 802.11n adapters are now featured in most new notebooks and even in many netbooks, and every enterprise-class vendor of Wi-Fi systems is offering 802.11n products today.

It's the Wi-Fi, stupid

The reason for this success is the approval in 2007 by the Wi-Fi Alliance of an interim interoperability specification for 802.11n, based on Draft 2 of the standard.

This specification is largely behind the perception that 802.11n was completed some time ago, because, after all, we ultimately buy Wi-Fi, not 802.11 anything. There are no conformance, compliance or compatibility tests for 802.11 per se; that role has been assumed by the Wi-Fi Alliance, whose many efforts have culminated in what is arguably the most successful trade association ever.

Indeed, after the torturous process that marked the early days of 802.11n, with political (and, to be fair, some technical as well) infighting among such long-forgotten vendor groups as MitMot, TGnSync, and WWiSE, it was essential for the Alliance to step in and bring some calm and order to what is the often-ugly process of standards creation.

And, as was the case with the original 802.11 standard, which also took seven years, the Wi-Fi Alliance's stamp of approval on an interim spec was the key enabling factor for market success.

As to the amount of time required for 802.11n to be created and finalized, well, the IEEE standards development process itself is designed to limit throughput, as it were, in the interest of reaching the best possible conclusion.

Keep in mind that a standard from a recognized standards-creation body, like the IEEE, has essentially the force of law. The IEEE has taken great pains to assure that standards development follows the principles of "due process, openness, consensus, and balance" and that the "right of appeal" is preserved at every step.

Complaints about the standards that result are indeed rare, other than, of course, regarding the amount of time involved to embody them. Most infighting is the result of either strongly-held technical positions, the desire to have one's intellectual property incorporated into the standard, or to preserve one's technical lead in the market.

The Wi-Fi Alliance has stated that its revised certification process based on the final standard will "preserve interoperability" with the current spec, again based on Draft 2.0 of the standard. With existing investment thus protected, we expect the progress of 802.11n towards all but certain dominance on a global basis will accelerate.

And this outcome was, of course, essentially enabled by the fact that few substantive changes were made between Draft 2.0 and the final standard. The most significant issue tackled was in enhancing coexistence with current 802.11-based products, along with lots of dotting of i's and crossing of t's.

System architecture and management move to center stage

Even the most conservative IT shops now have no reason left to avoid or postpone the adoption of 802.11n – the standard is done, the Wi-Fi Alliance has assured backwards compatibility, the performance of even today's pre-standard products is more than impressive, and prices for both infrastructure and adapters is more than reasonable.

Farpoint Group recommends that 802.11n be deployed only in the 5GHz. bands for most applications. A worry expressed by many potential customers has been with regard to both interference and coexistence in the well-subscribed 2.4GHz spectrum.

But there's really no reason to upset current 802.11g or 802.11b infrastructure, at least for the time being; the 5GHz. bands, initially used only for the sparsely-deployed 802.11a, have plenty of free spectrum for use of even the 40MHz. 802.11n channels that yield the highest performance.

We don't recommend operating 802.11n in the same channel with earlier technologies, even though this mode of operation is allowed in the standard. Performance for both technologies will suffer, and, again, there's usually a lot of otherwise unoccupied spectrum above 5GHz. for greenfield 802.11n buildouts.

Next steps

For vendors, 802.11n is now part of the jacks-or-better required to be in the WLAN game. Going forward, the real focus will shift to the remainder of the system, not the radio.

Recent announcements from the major players have, for example, focused on the control plane, which can be thought of as the operating system of enterprise-class WLAN infrastructure and the element where traffic-control policy is implemented.

Advances in traffic scheduling are yielding significant improvements in performance irrespective of the particular radio technology employed. Traffic analysis, deriving from the mathematical field of queuing and scheduling theory, has been employed in operating systems for years, with very positive results. So knowing what to move, and when to move it, is likely to provide the next performance boost beyond 802.11n alone.

While some have argued that overprovisioning should be sufficient to assure the throughput and responsiveness required, this approach ignores the fact that demands on networks only grow over time, in terms of not just throughput, but also number of clients, data object size, duty cycles, and requirements for time-boundedness. Yes, 802.11n provides a significant increase in capacity, but it unlikely to be sufficient in and of itself to meet most enterprise requirements.

The remainder of the infrastructure, in both the control and management planes, is likely to be the key differentiator from this point forward, rather than specific 802.11n radios.

So where do we go from here?

The IEEE's 802.11 group is now hard at work on even more additions to the standard, including efforts to further boost throughput. While 802.11n will offer throughput potentially much greater than the 100Mbps still common in many enterprise settings, it cannot achieve anything close to gigabit Ethernet.

For this reason, 802.11 has formed two new task groups to specify wireless links in excess of 1Gbps: 802.11ac is using spectrum below 6GHz., and 802.11ad is looking at spectrum around 60GHz. Multi-gigabit throughput for both is a real possibility, although there are issues with range and propagation at millimeter-wave frequencies.

So while gigabit Ethernet is a more than desirable option for interconnecting 802.11n access points, it will very likely see its role as a primary access technology challenged by future 802.11 developments.

The standards group is also continuing work on such diverse areas as meshes (802.11s), management (802.11v), management-frame security (802.11w), video streaming (802.11aa), to name but a few. Similarly, the Wi-Fi Alliance is hard at work on such areas as wireless network management, certifying some optional features, enterprise-grade enhancements to their Voice Certification, security enhancements to address management frames (corresponding to 802.11w), and a very significant program for device-to-device connectivity.

The opportunities at this point appear unlimited. 802.11n will, we believe, play a major role in the future of cellular networks, as the carriers increasingly rely on Wi-Fi to offload both voice and data on their crowded (and expensive) spectrum. 802.11n will become a fixture in both dual-mode cellular and dedicated cordless handsets over the next few years. Wi-Fi is even making inroads as a personal-area network technology.

With the ratification of 802.11n, WLANs have taken another big step forward towards ubiquity across enterprises, applications, and, for that matter, the entire planet.

Mathias is a principal at Farpoint Group, a wireless advisory firm in Ashland, Mass. He can be reached at craig@farpointgroup.com.

This story, "Wi-Fi eyes global domination" was originally published by Network World.

Copyright © 2009 IDG Communications, Inc.

Bing’s AI chatbot came to work for me. I had to fire it.
Shop Tech Products at Amazon