Skip Links

Broadcom: Making the Case for 802.11ac

Many (OK, most) assume that 802.11ac is the anointed successor to 802.11n. A recent conversation with Broadcom finds a convincing case here – with a couple of caveats.

By Craig Mathias on Wed, 12/14/11 - 2:10pm.

Last week, I spoke with a senior manager at Broadcom, a company with very deep roots indeed in wireless LANs, and one of the leading suppliers of Wi-Fi chips. The purpose of the call was to explore 802.11ac, the "gigabit-wireless" standard now under development that will operate in the 5 GHz. bands. You will recall that the other standard here, 802.11ad, operates at 60 GHz. I'm not in the majority with this opinion, to be sure, but I believe that there is a significant role for .11ad in enterprise WLANs, in addition to 60 GHz. serving as a logical homeland for in-room HDMI replacements like the Zyxel AeroBeam link (which is not, BTW, based on 802.11d) we recently reviewed. But I think most people believe that .11ac is the logical longer-term replacement for the venerable and hardly-dead-yet 802.11n.

First off, that statement is absolutely correct at face value. .11ac will be backwards-compatible with .11n, so all of us will be buying only .11ac products within three years (as .11n production slows), and a lot of these will operate throughout their entire useful lives in .11n mode. I am expecting something of a "better n than n" effect here, due to advances in implementation technologies and radio architecture, but that advantage will be likely only a few percent. No matter - we'll first see .11ac clients working with good old 300 (and some 450) Mbps .11n, which will remain in place in many venues for the long haul. Residential customers will be the first to see full-speed .11ac, but connecting to even 35 Mbps (as in my case) ISP links will leave many asking "what's the point?"

It gets a little worse. Even in full .11ac mode, performance will typically be on the order of up to 433 Mbps for handsets (1x1, 80 MHz. channel), and around 867 Mbps (again, peak) assuming 2x2 and again that 80 MHz. channel. The 80 MHz. channel is of some concern, as (a) that's a lot of real estate and finding a relatively unused contiguous chunk that big will prove increasingly difficult, and (b) the regulators will undoubtedly require strict limits on transmit power, with consequential limits on range and rate-vs.-range at least to some degree. Of course, we won't know the real story here until products are available, and I expect, as was the case with .11n, that we'll see some of those well before the standard is finished and the ink on the Wi-Fi spec is dry - as early as the second half of this year. But, remember, perhaps here more than ever, numbers like 433 and 867 are the rates that the marketing departments out there will absolutely guarantee that their products will never exceed.

There is a big upside as well - the standard will allow aggregate throughput on a given 160-MHz. channel (which may be implemented as 80x2) to reach as high as 6.93 Gbps! I don't expect we'll see this kind of performance anytime soon, just as has been the case with 4x4 .11n, but it's nice to know that the standards guys and girls are thinking that far ahead. But if you're thinking ahead, as I'm sure you are, even given all of this you still won't rule out solutions based on 802.11ad - at least not yet. Seven GHz. of virgin spectrum? C'mon!