Continuing in the 802.11ac wave 2 series where I dissect each feature in the “wave 2” grab-bag of technologies, this post focuses on the history of beamforming.
Beamforming was first standardized in 802.11n, but it did not become part of the standard feature set in commercial products. In part, that was due to a plethora of options. (Please, no comments on the persistence of standards; I’ve heard it already, and it’s wrong)
Fundamentally, beamforming concentrates energy, which improves signal-to-noise ratio, and that can improve the data rate you see at the Wi-Fi client.
The baseline scenario for most Wi-Fi APs is that they use omnidirectional antennas. As the name implies, an omni antenna sends energy in all directions. For many scenarios, you don’t know where the receiver is, so spraying energy out in all directions is the right thing to do.
The problem with using an omni to reach a client is that you don’t know where it is. To borrow an old saying from the advertising world, with an omni, half of your transmit power is wasted, but you don’t know which half. (Though, to be fair, ad agency giant WPP thinks the saying is apocryphal.
In concept, beamforming is a magical idea, and it promises something for seemingly nothing. Learn where the client is, and rather than send our transmit power everywhere, we send it only where the client is. More power means more signal to rise above the muck of the noise floor. Better signal-to-noise ratios (SNR) means a higher data rate, and we’ve improved the performance of the link without needing to do anything other than make sure packets go in the right direction.
(Figure 4-1 from 802.11ac: A Survival Guide)
If beamforming is such a good idea, why didn’t we start using it right away in 11n?
The high-level view is that it’s hard! While it’s easy for us as humans to look at the spatial relationship between two devices and see what direction the AP needs to fire off packets to the receiver, it’s much harder in real life. The complexity of the radio channel requires that a beamformer needs to understand in a much higher level of detail how to send frames so that they travel primarily in the intended direction. In the picture above, as the AP, how do you know how to create a transmission that preferentially goes into the green wedges.
And of course, you can’t get it wrong. Beamforming increases the energy available in some places, and the goal is to make sure that the “hot spots” of a beamformed frame are where the receivers are. It’s like squeezing on a balloon. If you decide the balloon is going to expand along one direction, it’s going to contract along others. Or, returning to the hot spot analogy, there are no hot spots without creating cold spots.
To successfully shift the energy around a transmitter’s area, you need to understand what the radio channel looks like, and that means you have some sort of measurement system. 802.11n specified multiple options, which meant chaos in the marketplace because both sides needed to negotiate using the same option. And when most devices supported incompatible options, nothing happened. Instead, beamforming in 802.11n used a less-accurate method of channel assessment by looking carefully at the received acknowledgements to make educated guesses about the state of the channel.
If beamforming in 802.11n was not widely implemented, what changed in 802.11ac? And the answer is simple: 802.11ac selected one method of channel measurement (called Null Data Packets or NDPs), so any 802.11ac beamforming device is be able to work with any other 802.11ac beamforming device.
With the first wave of 802.11ac products, we’ve seen more beamforming than we did in the entirety of the 802.11n era.
~~~
Other posts in this series on 802.11ac wave 2:
802.11ac wave 2: Regulations around 160 MHz channels (By Matthew Gast)
160 MHz channels in 802.11ac wave 2: facts and fictions (By Matthew Gast)
What does 802.11ac "wave 2" mean and do you need it? (By Matthew Gast)
Wait on 802.11ac Wave 2 until mobile devices catch up (by Metka Dragos)
Deploying 802.11ac won't force a wired network upgrade (by Metka Dragos)