As I mentioned before, I'm at Interop New York all this week, where I once again have the privilege of Chairing the Mobile Business element of the event. The best part about attending a conference often isn't the conference sessions or even all the cool new products on the exhibits floor. It's the discussions with attendees and speakers that (most often) occur in the hallways. And I spent almost all of yesterday doing just that - having one interesting conversation after another on topics ranging from mobile/mobile convergence (Wi-Fi is better than femotcells) to mobile WiMAX (in very deep yogurt that might in fact more closely resemble the La Brea Tar Pits), to antennas (MIMO antennas in handsets are a challenge, but there are some brilliant ideas here, as we heard in one session from SkyCross), to asynchronous vs. synchronous systems (a topic for another day), to why CDMA won the 3G race (another interesting bit of history), to whether we really need 4x4 MIMO implementations of 802.11n. OK, on that last one, a few words starting with - duh, yes.
Just for background, 4x4 means four transmitters and four receivers, required under the 802.11n standard to get the full 600 Mbps possible. As I have said many times, throughput on wireless isn't really about throughput - it's about capacity. If I get my bits on and off the air faster and more reliably (so that the need for retransmissions is minimized), there's a lot more capacity left for you. Serial communications channels (like Wi-Fi) thus benefit from MIMO in a number of dimensions - most notably throughput, capacity, and reliability, but also range, important in many applications, with the optimization of these almost always a good thing (the exception being, again, range, and again depending upon application). So even if one doesn't strictly need 600 Mbps, one benefits from its provisioning regardless.
The downside of 4x4 is potentially cost, a small degree of added complexity, possibly power consumption and circuit board real estate, and, most notably at present, availability. The only two vendors I know of building 4x4 chips today are Quantenna and Qualcomm, and I don't know of any 4x4 end-user products currently available, although I do expect this to be corrected shortly. But I'm willing to predict that this area will see significant activity over the next few years and indeed become quite competitive, especially for fixed applications like video distribution in the home. Note this brings up the whole subject of compressed vs. uncompressed video streaming, which, being a hugely controversial topic all by itself, I'll return to in 2010. For now though, yes, I know - 1x1 and 3x3 are just getting established, and 4x4 might seem exotic. It's not. And you're going to be hearing a lot more about it over the next few years. And, finally, I wouldn't be surprised if 4x4 catches fire pretty quickly. You might even own such a product yourself, yes, in 2010.
Finally, Interop New York is surprisingly busy this year. I must confess I was a bit worried, given the economy and all. But I don't think "pent up" adequately describes the demand we'll see next year, as the recession ends, because - well, we're really, really tired of the "Great Recession" now, and I think, apart from the profit motive, of course, we're all really happiest when we're working and busy. So, if you've been concerned that you're underemployed (or worse, as has sadly been the case with so many for the past year), fear not - barring any totally boneheaded move on the part of the government (which we, again, sadly, can never entirely bar regardless of the macroeconomic climate), I think we're about to turn the corner, and activity in networking and IT in general is about to see a marked improvement.
Mathias is a principal at Farpoint Group, a wireless advisory firm in Ashland, Mass.