One may ask why it took two years for the Institute of Electrical and Electronics Engineers (IEEE) 802.11 working group to produce the final standard of 802.11n, when an interoperable and practicable Draft 2.0 was available back in 2007. The answer involves peeling back the layers of the operation of the IEEE standards development organization.
IEEE itself is a broad organization, known mostly for its academic work, publishing journals and organizing interest groups, and serving as an organization of individual member engineers. But, for networking, the IEEE branch that develops standards is in charge, creating and driving for the international adoption of particular technologies, that often go by the more friendly terms Ethernet (802.3), Wi-Fi (802.11), ZigBee (802.15.4), and WiMAX (802.16e), to name a few. Within each particular working group–802.11, for example–there are task groups that produce specific amendments. 802.11n is one of those amendments.
802.11n started back in September 11, 2003, as an official task group, chartered with the goal of producing at least 100Mbps of throughput, with the goal clearly that of taking on the wireline rival of 100Mbps Fast Ethernet. The initial process of almost every 802.11 task group is to take a couple of years to decide what technologies will be used to accomplish the goals of the group. Now, going into the task group, there are always ideas on how to proceed. Some task groups are very high-level, such as 802.11v, and so end up with a smattering of different solutions to different problems. But the ones that fundamentally alter the way the radios themselves work have far more concrete algorithms and signal types in mind when they start. 802.11n would have to use multiple streams of data–MIMO–to get to its 100Mbps number without changing the channel widths.
For the first few years of 802.11n, there was vigorous debate of exactly what new ideas in increasing throughput should be taken. Right away, proposals included aggregating packets into wireless “jumbo” frames, allowing for bonded (40MHz wide) channels, and improving the efficiency of the signal encoding down to the bit level. But different flavors of these existed, resulting in competing proposals. These competing proposals had to be weeded out to produce a first draft–Draft 1.0. And, after a fair amount of discussion both in and out of IEEE, a compromise solution was reached and became that draft. Draft 1.0, produced March 19, 2006, was not something a person could go out and build to. The broad brush strokes were there, but major areas of the canvas were left blank, in a desire to get something concrete together. Therefore, 802.11n took its longest gap between drafts by far–one and a half years–to fill in the details and produce a draft that someone could build to. This became Draft 2.0, produced in February 2007.
At the same time, the same industry players were coordinating through the Wi-Fi Alliance to make sure that there was a certification program for this draft, knowing that this draft represented something that could be built and sold with low risk. The Wi-Fi Alliance then produced its 802.11n Draft 2.0 certification program.
So, why the two year delay from 2.0 to now? There are two reasons for that. 802.11n is one of the most successful programs in 802.11, having come together so quickly. Most other Draft 2.0s in 802.11 are not really usable, and change tremendously. But 802.11n would not. This meant that the pressure was off, so to speak, and the group could focus on improving the text itself, as well as pursuing fixes to optional features that are not meant to be common in enterprise or consumer 802.11 deployments. In fact, 802.11n has roughly produced a new draft every three months, capturing whatever changes were made at one of the six annual 802.11 meetings and voted upon by the membership in between. As a large international organization, IEEE has its review processes, and these processes involve continual republication and voting, until as large of a consensus as possible has been built–90%+ approval is quite the norm.
As a result, although it may seem that 802.11n took quite a long road–and for those involved, it was not a short process from start to finish–the development of 802.11n was quite methodical, and more importantly, reached a point of interoperability far quicker than most anything that comes out of international organizations. Given the size of 802.11n–560 pages–this is quite an amazing thing.
