Matthew Gast is director of product management at the Wi-Fi firm Aerohive Networks. He is also chair of the wireless network management task group at the Wi-Fi Alliance, an industry body devoted to ensuring interoperability.
Gast also works on Wi-Fi standards at the IEEE, and has been involved in the development of the successors to the 802.11n, the latest commercial implementation of Wi-Fi that has become familiar in laptops, smartphones and routers.
The next two important flavours of Wi-Fi will be known as 802.11ac and 801.11ad. Both are intended to offer speeds of up to 1Gbps, although the two have different uses. ZDNet UK caught up with Gast at the IP Expo in London to find out what to expect from gigabit Wi-Fi, and when to expect it.
Q: 802.11n is already pretty fast. What should we expect from future versions of Wi-Fi?
A: We delivered 600Mbps in 802.11n, and we're waiting for the hardware engineers to catch up. Six hundred is a big number but it's not nice and round — a gigabit is nice and round. 802.11ac and 802.11ad have the same goal, which is to be really fast. The major difference is the frequency they operate in. 802.11ac is gigabit at less than 6GHz, so that's very evolutionary. 802.11ad is intended to go at 60GHz — there's some new spectrum there.
If 802.11ac is evolutionary, what's the use case for 802.11ad?
You can build very high-density networks — distance is not always a plus. Think of a university lecture hall. We're spending a lot of time trying to engineer networks that work well; you need a lot of access points for capacity, but you need to make sure they co-operate.
Do you mean mesh networking?
Possibly — with a bunch of short-range access points, you might not want to put wires to all of them, so some sort of self-organising mesh could be one way to connect them. Because we're so early [on the development], a lot of this conversation is really speculative. Another potential use case would be to build a multiband access point — we have dual band now. If you [provided the] service with 802.11ac and then you used 802.11ad for backhaul, you could build a self-organising mesh. At 60GHz, there's much less deployed there [than in the 2.1GHz spectrum currently used for Wi-Fi]. It's less cluttered. You might be able to have antennas that adapt to it, or it might be the other way round — you might do mesh and access at the same time on the same radio.
What's the timescale for 802.11ac and 802.11ad?
We got a model with 802.11n that was pretty good in terms of how it rolled out. The second draft was good enough to have pretty good interoperability with the pieces people cared about, and that's how we got to 100Mbps.
Wasn't the standardisation process for 802.11n something of a mess? The industry put its products out before the standard was ratified...
The industry also learned from that experience. Interoperability is also important — that's part of the reason the second draft of 802.11n was what wound up being in the complete standard. We've learned the message. You learn a lot by building hardware.
Regarding 802.11ac and 802.11ad, my rough guess is that we'd be talking about late next year or early 2012 before we get chips available. Once the chip is available, we can start trying to build hardware based on it.
Why would people need these new standards?
802.11n made it possible to replace existing copper networks [with Wi-Fi]. You had to have a reason to use 802.11a/b/g — you had to really want mobility, because you had to trade off speed for mobility. Now, a lot of people don't have landline networks anymore; they grew up with mobiles. 802.11n is enabling the same sort of transition in networking.
Looking back on it, the MacBook Air was an incredibly significant device, because Apple said 'We're building this laptop; we don't need to put in an Ethernet port; it's good enough to rely on the wireless connection'. It was visionary of them to do that, but 802.11n has caught up.
Wouldn't it still be true to say, though, that 802.11n remains overkill for many people?
It depends on what you're doing. For instance, health care, moving medical records around with detailed images — you want someone to look at an image on their iPad and offer a diagnosis. Or if you want to do something like voice, 802.11n is very helpful for voice [due to its low latency]. Also [managing airtime usage by difference devices] — by scheduling very carefully it was possible to increase the overall capacity of the network by essentially rewarding the devices which were fastest.
What will 802.11ac and 802.11ad do, aside from being very fast?
802.11ac will use multi-user Mimo. With beamforming [as used in 802.11n], if I were an access point, I could send my transmission at you and focus the direction in which I transmit. Because radio waves add, I can do both at the same time, and send out in different directions from the same set of antennas. The access point becomes like a switch, transmitting to two multiple stations at the same time. So you have something that's faster and can download multiple frames at the same time.