All Things 802.11ac - Question 5: What is Wave 2?

  • 1
  • Question
  • Updated 3 years ago
  • Answered
There’s a lot of talk (and already some here today) about “Wave 2” of 802.11ac (I am, BTW, not even sure that’s an official term, but everyone seems to be using it regardless). No matter - what is Wave 2? Is it just about multi-user MIMO (MU-MIMO), or are there other associated features and benefits? When will we see it in the market? And should IT departments wait for Wave 2 before deploying any 802.11ac solutions?
Photo of Craig Mathias

Craig Mathias

  • 63 Posts
  • 0 Reply Likes

Posted 3 years ago

  • 1
Photo of Matthew Gast

Matthew Gast, Director of Product Management

  • 284 Posts
  • 60 Reply Likes
I'd say that it's a semi-official term. We have a long experience in the industry with doing phased solutions -- WPA1 and WPA2, 802.11n draft 2 and final, and now 802.11ac draft and, I assume, in the future, a final 11ac certification plan.

802.11n was the first spec that was so complex that it couldn't come out all at once. Chip vendors focused on dividing up the specification into a few distinct sets of functionality, and delivered them in phases. I'm not sure why we started calling them "waves," but I guess that's the new word.

There are three potential features that are likely to be part of wave 2 are (1) 160 MHz channels, (2) a fourth spatial stream, and (3) widespread beamforming, perhaps with multi-user MIMO.

There's definite benefits to another spatial stream. Beamforming and MU-MIMO are exciting, and have a great deal of potential, but they'll be tricky to implement.
Photo of Craig Mathias

Craig Mathias

  • 63 Posts
  • 0 Reply Likes
How does MU-MIMO work, and do you think it will really be effect? Is it worth waiting for?
Photo of Craig Mathias

Craig Mathias

  • 63 Posts
  • 0 Reply Likes
Make that "effective"
Photo of Nick Lowe

Nick Lowe, Official Rep

  • 2094 Posts
  • 360 Reply Likes
As Matthew says, it's simply:

1) MU-MIMO
2) More Spatial Streams
3) Wider Channels
4) Generational Iterative Improvements - Power Consumption, Signal Processing
Photo of Matthew Gast

Matthew Gast, Director of Product Management

  • 284 Posts
  • 60 Reply Likes
Of those components, MU-MIMO is definitely the wildcard. We know what we'll get out of 160 MHz channels (another 2x in max speed) and a fourth spatial stream (1.33x in max speed). Much of the range in quoted wave 2 performance is due to differing predictions on how much speed MU-MIMO will add.
Photo of Nick Lowe

Nick Lowe, Official Rep

  • 2094 Posts
  • 360 Reply Likes
It would be great to see in the final WiFi Cerification for 802.11ac a requirement for manufacturers to specify the nxn:n capabilities of whatever they're selling in the product sheet, be it an laptop, tablet or AP.

It has been a rather large frustration with 802.11n that this hasn't been the case.

It would certainly help with Wave 2 products and differentiating them.
Photo of Craig Mathias

Craig Mathias

  • 63 Posts
  • 0 Reply Likes
Couldn't agree more! Perhaps we need a federal law here...
Photo of Matthew Gast

Matthew Gast, Director of Product Management

  • 284 Posts
  • 60 Reply Likes
There's a basic picture of how MU-MIMO works over in this thread:
http://community.aerohive.com/aerohiv...

Here's a slightly more complicated version of the same figure. Each of the colors remains the same as in the basic picture, but with the addition of mathematical explanations. Q is the steering matrix, which roughly tells you how to direct the energy to a particular receiver. H is the channel matrix, which describes how the channel responds. The product of the two matrices tells you what the receiver gets to see.



For a video explanation, check out the video of my Wireless Field Day 5 presentation, which you can get to from our wrap-up on the event: http://blogs.aerohive.com/blog/the-en...
Photo of Matthew Gast

Matthew Gast, Director of Product Management

  • 284 Posts
  • 60 Reply Likes
One important note about the figure above: the ideal situation is that the steering matrix produces a signal for its intended receiver ("max" in the figure) while producing nothing for other receivers ("null" in the figure). The number of antennas (technically, independent transmission chains) on the AP helps you do this null-steering to prevent interference between clients.
Photo of Nick Lowe

Nick Lowe, Official Rep

  • 2094 Posts
  • 360 Reply Likes
I assume this steering matrix and its computational complexity is the reason we're starting to see co-processors on 802.11ac cards.
Photo of Craig Mathias

Craig Mathias

  • 63 Posts
  • 0 Reply Likes
You raise a good point - this stuff is computationally intensive. What about battery life?
Photo of Matthew Gast

Matthew Gast, Director of Product Management

  • 284 Posts
  • 60 Reply Likes
One of the reasons for having CPUs (or CPU cores) on the interface is that it serves as a pre-processor to help out the main system CPU. Ethernet cards went to using specialized ASICs for the same reason. Most of the CPU cores are ARMs, which are fortunately naturally pretty low-power.
Photo of Nick Lowe

Nick Lowe, Official Rep

  • 2094 Posts
  • 360 Reply Likes
... Which makes it a net win for power consumption if you've not pushed it to the main CPU. Still, there has to be an overhead vs. not doing it at all. Guess it's all about trade-offs.