All Things 802.11ac - Question 2: What is Beamforming?

  • 2
  • Question
  • Updated 5 years ago
  • Answered
I think most people know that beamforming enables a transmitter to focus energy on an intended recipient, improving range, throughput, and reliability. But I also think most people don’t know that beamforming is a standard feature in 802.11ac. What are the real-world benefits of beamforming, and are there any issues with its implementation (and I’m assuming some implementations will be better than others...)?
Photo of Craig Mathias

Craig Mathias

  • 63 Posts
  • 0 Reply Likes

Posted 5 years ago

  • 2
Photo of Nick Lowe

Nick Lowe, Official Rep

  • 2491 Posts
  • 451 Reply Likes
Are the 802.11n Atheros chips used in Aerohive's APs today capable of this standards based beamforming with a software update?

For example, the AR9390 in the HiveAP 330.

Or do we have to purchase 802.11ac APs to get this because of fundamental changes that require a refresh in hardware?
Photo of Matthew Gast

Matthew Gast

  • 284 Posts
  • 63 Reply Likes
The biggest issue with beamforming prior to 802.11n was that the standard had multiple options for beamforming. It defined a negotiation protocol that enabled the AP and client to choose a mutually-compatible beamforming method, but in practice, there was not much interoperability. Each device supported one method, but many methods were supported by only one chipset vendor. As a result, it wasn't worth turning on beamforming in 802.11n because it would only benefit a few clients.

802.11ac simplified beamforming by settling on just one method, called Null Data Packets (NDPs). With everybody agreed on using NDP-based beamforming, interoperability is much easier to achieve.
Photo of Nick Lowe

Nick Lowe, Official Rep

  • 2491 Posts
  • 451 Reply Likes
That's a very fair point, and you would also require client driver updates to get it working properly, which is unlikely to see wide deployment.

(We're seeing this problem today in another sense with 802.11r and the updated RSN IE that has an additional AKM and drivers in the field not playing ball with it.)
Photo of Sami Mousa

Sami Mousa

  • 1 Post
  • 0 Reply Likes
WHEN SOMONE USE rsN OR akm CAN YOU EXPLAIN WHAT STAND FOR PLEASE?
Photo of Nick Lowe

Nick Lowe, Official Rep

  • 2491 Posts
  • 451 Reply Likes
Matthew can best comment on this but think of RSN as being the working name that became WPA2.
Photo of Matthew Gast

Matthew Gast

  • 284 Posts
  • 63 Reply Likes
RSN is the "Robust Security Network." It refers to the information element that describes security settings to receivers. AKM is the "authentication and key management suite" that describes how you are performing authentication (802.1X, pre-shared key, etc). See clause 8.4.2.27 in 802.11-2012, which you can download for free from the IEEE.
Photo of Nick Lowe

Nick Lowe, Official Rep

  • 2491 Posts
  • 451 Reply Likes
Photo of Matthew Gast

Matthew Gast

  • 284 Posts
  • 63 Reply Likes
The real-world benefit of beamforming is probably an increase in SNR of about 3 dB at best. The current FCC and ETSI rules require that beamformed transmissions reduce their transmit power by either 3 or 4.8 dB, so whether you can achieve higher performance with beamforming under the current regulations is a far more interesting question. I wrote about this in the spring on the Aerohive blog: http://blogs.aerohive.com/blog/the-wi...
Photo of Craig Mathias

Craig Mathias

  • 63 Posts
  • 0 Reply Likes
OK - but how will it work in practice? It would seem that having beamforming only on the transmit side gets you 90% of the way there, as the signal is inherently more reliable.
Photo of Matthew Gast

Matthew Gast

  • 284 Posts
  • 63 Reply Likes
I should have been clearer about the difference between explicit and implicit beamforming. Implicit beamforming uses things like lost ACKs to figure out how to steer a transmission, and some implicit beamforming implementations have only minimal performance gains. Explicit beamforming is based on measuring the channel, and it requires cooperation on both sides to collect the channel state.
Photo of Jake Snyder

Jake Snyder

  • 3 Posts
  • 0 Reply Likes
How are the benefits of beamforming affected by distance? Is this a small cell only technology, or does it provide benefit in larger cell deployments?
Photo of Tony Boudro

Tony Boudro

  • 6 Posts
  • 2 Reply Likes
I'm with Jake...large manufacture setup, open but distanced.
Photo of Matthew Gast

Matthew Gast

  • 284 Posts
  • 63 Reply Likes
Beamforming is a mid-range technology. At short distances, you're already at top speed. At long distances, you need to transmit the channel measurement packets and they eat up enough time the overall transmission (measurement + data) takes longer than just using the slower speed. It's great in the middle ranges, where you can spend time measuring the channel and there's a serious return of a data rate or two.
Photo of Craig Mathias

Craig Mathias

  • 63 Posts
  • 0 Reply Likes
Still, 3 dB is nothing to sneeze at. Coupling this with even a slight bump up in deployment density, which will likely regardless be required given increasing client density, overall throughput and reliability should benefit., Still, I worry about interoperability and the highly-varying quality of implementations, which we saw with beamforming in the pre-standard era. Beamforming remains a difficult technology to master...
Photo of Matthew Gast

Matthew Gast

  • 284 Posts
  • 63 Reply Likes
Oh, yes, 3 dB is nothing to sneeze at, but it's a best-case scenario. Transmit power rules require that power be reduced by 3 dB for a two-antenna transmitter or 4.8 dB for a three-antenna transmitter. So, if I have to drop my power by 3 dB to use beamforming, and will get back a 3 dB gain in the best case, beamforming may not be feasible until we can get the rules changed.
Photo of cmd1775

cmd1775

  • 2 Posts
  • 0 Reply Likes
No beamforming, no MU-MIMO, correct? Don't the benefits of MU-MIMO make beamforming a win, even with the FCC handcuffing it (for now)?
Photo of Matthew Gast

Matthew Gast

  • 284 Posts
  • 63 Reply Likes
MU-MIMO is a complicated subject. First-generation beamforming -- which is all that's on the market today and for the foreseeable future (probably until next year at the earliest) -- is all single-user. It's important technology for us to develop, but it's not a big win until we can use it for multi-user transmission.

Even in multi-user transmission, it might not always make sense. There will be some degree of inter-client interference, so MU-MIMO trades peak speed to one device for an improvement in the total overall network throughput.