A few weeks ago, Qualcomm was demoing it’s Wilocity based 802.11ad 60Ghz Wi-Fi silicon for 4K streaming. For the most part it works quite well with one major caveat that could cause some headaches among the unwashed masses.
The 802.11ad technology is nothing new nor is Wilocity’s implementation, there are a bunch of devices with it on the market already. We first showed you a few of these in mid-2013 including a Dell solution that effectively streams video like Qualcomm was showing recently. While the Displaylink/Dell solution wasn’t streaming 4K at the time, Qualcomm hadn’t purchased the company yet either, not that these two things are in any way related.
The demo was pretty simple, stream a 4K video over 802.11ad and run a benchmark to show how much usable bandwidth was left over. 802.11ad uses the 60Ghz band and is theoretically capable of transferring 3Gbps. Just for the record, the 802.11ad phy rate is 4.6Gbps so that 3Gbps number should actually be achievable. The demo showed this through two Snapdragon 810 reference designs, each with a Wilocity modem attached via USB.
The little red disc is the 802.11ad radio
The demo wasn’t much to look at but it did work. The 4K video was sent from the tablet in the foreground to the one in back which played it on a 4K monitor above via HDMI. There was also the Iperf bandwidth benchmark running to measure the remaining throughput. In this case that number hovered around the 180MBps/1.6Gbps mark and the video of course did not waver. In short it did what Qualcomm claimed it would.
Small 802.11ad modules with a big baseband board
There was also a few of the bare 802.11ad modules on display with four rectangular radios and one squarish one that included the baseband. The smaller radios had up to 32 antenna inputs, the larger one only had six. That brings us to the caveat we mentioned earlier, that is the 60Ghz band means 802.11ad is effectively a line of sight (LoS) technology.
We called line of sight a caveat because it is not a good or bad thing, it just is. On the up side it can be used in dense areas like apartment buildings or offices without stepping on the user next door. To be fair it may be restricted to every other person and cube farms are pretty much out, those walls are too thin to block a 60GHz signal by themselves They may drop the signal quite a bit but not enough to avoid problems. This is the good part of LoS.
The bad part of LoS is why you might actually need all 32 of those antennas in a device, a cube wall might not be enough to stop an 802.11ad signal but your head is much thicker than most cube walls. Your co-workers usually have even thicker heads too, and bosses, well they will definitely block the signal. In any case the orientation of the devices and what is between them matter a lot in this case so a moving 802.11ad radio is a challenge. This is the down side and it is a pretty small one for things like the Dell docking station but can be much more problematic in a handset.
This all brings up the biggest problem facing 802.11ad, users and stupidity. Most of you reading this will get the tech behind the LoS ‘problem’, most purchasers won’t, they will just get frustrated. The biggest problem as I see it is in the name Wi-Fi. If average consumers can read the box without ruining it by drooling, they will see Wi-Fi and expect a few things like a range in the hundreds of feet and the ability to go through walls. They will try 802.11ad and when it doesn’t do that, problems will ensue. These problems could be high return rates, increased tech support costs, and the like. The industry needs to do more on the nomenclature and education front to avoid this pitfall.
One potential solution is called FST or Fast Session Transfer. FST is a way to seamlessly hand off sessions from 2.4<->5<->60Ghz without interrupting the data. Since it is part of the 802.11ac base specification it should be widespread in a few years and most 802.11ad implementations should have it built-in. Hopefully. 2.4Ghz legacy devices on the other hand…. In any case this may work for a cell device which has the gateway and ‘other side’ taken care of, for everything else as they say, your results may vary. Unless the industry makes it very clear what 802.11ad does and does not do, the fact that it works quite well will be irrelevant.S|A
Latest posts by Charlie Demerjian (see all)
- More on Intel’s 10nm process problems - Sep 17, 2018
- Intel puts out another 14nm 2020 server platform - Sep 11, 2018
- Why Can’t Intel Supply Enough 14nm Xeons? - Sep 10, 2018
- Intel can’t supply 14nm Xeons, HPE directly recommends AMD Epyc - Sep 7, 2018
- AMD reintroduces the Athlon name with two CPUs - Sep 6, 2018