Wednesday, November 18, 2009

Wireless High Def Video: Do We Have a Winner?

Nope, not so much.

But we’re closer than ever before.

News broke yesterday that Sony has dumped Amimon’s 5 GHz approach in favor of SiBEAM’s 60 GHz solution. While I’m still not ready to call a winner in this race, SiBEAM and their WirelessHD consortium are now firmly ensconced in the catbird seat. Adding another marquee name like Sony to announced customers LG and Panasonic obviously bodes well for the WirelessHD camp. And, despite the fact that worldwide 60 GHz harmonization efforts are anything but final, I think we’ve finally reached a confidence level where consumers can purchase a solution based on silicon from a company who’s going to be around for awhile, and have a product which will work long-term.

While both Amimon and SiBEAM have spent a ton of money forming their alliances (WHDI and WirelessHD, respectively) and working the standards bodies, Amimon’s choice of spectrum location has long been a concern to many of us in the industry. Earlier this decade, one thing we had going for us at Tzero was that our ultra-wideband wireless high definition offering lived in a spectrum slot (3.1-4.8 GHz) with little potential for interference. Plus, UWB’s 528 MHz channel bandwidth ensured that we’d have a whole lot of capacity in case we did in fact bump into interferers. Finally, the fact that we were WiMedia Alliance certified meant that we’d behave properly with Certified Wireless USB solutions we might encounter.

What killed us, of course, is the ridiculously low power output allotted to UWB by the FCC. I’m not faulting the FCC here—simply stating the fact that trying to take a technology designed for very close proximity cable replacement (for wired USB-attached peripherals) and adapting it to carriage of very high bitrate wireless video across a living room was one helluva challenge. To illustrate—put your hands in front of your body, separated vertically by as much as your arms could stretch. Then look at the width of your hands themselves. The vertical distance is the transmission power of Wi-Fi; the horizontal distance (your hand width) is the channel bandwidth (20 MHz with 802.11a/b/g, 40 MHz with 802.11n).

Now, rotate your arms 90 degrees. The horizontal separation is the 528 MHz channel bandwidth of UWB; the height of your finger (go ahead, pick the middle one) is UWB’s transmission power. Would UWB have made a great solution for transmitting high def video wirelessly? Sure…if you had the box close enough to the TV that the signal stayed robust. You know, within a few feet. Like, within the range of a one or two meter HDMI cable.

Which pretty much defeats the entire friggin’ purpose.

While Tzero was marching towards irrelevance, Amimon and SiBEAM were unveiling their solutions for delivering HD video wirelessly. The key capability is that both companies’ approaches claimed and demonstrated robust in-room (or beyond) capabilities—and they actually delivered. Ten months ago at CES 2009, I met with all the relevant players in the space, including Amimon and SiBEAM. I was extremely impressed with SiBEAM’s solution for in-room connectivity, although by their own admission they were not suitable for multi-room use. Amimon’s demo in a four-room suite at the Las Vegas Hilton showed both in-room and whole-home connectivity, which was mega-impressive. Playing a fast-twitch first person shooter game over the Amimon link felt like I was connected via a wire—no perceptible latency. None.

However, I left CES with the same nagging feeling that had bugged me ever since Amimon entered the market—that a 20 or 40 MHz channel in spectrum soon to be occupied by a significant number of high-band 802.11n devices was a recipe for disaster. Did Amimon demo well in 2009? Yes, absolutely. But, we’re on the cusp of dual-band 802.11n devices making their inexorable march from the ridiculously crowded 2.4 GHz band to the currently-light-but-soon-to-be-saturated 5 GHz bands; sure, 11n at 5 GHz has a whole bunch more channels, but I just couldn’t reconcile Amimon’s pitch with real-world performance. Having suffered through stupidly high bit-error rates on my own 2.4 GHz wireless networks, I was and remain hugely concerned that 11n devices are going to wreak havoc on Amimon-based solutions. Does Amimon have a solution for detecting and minimizing channel conflict? Sure, you betcha.

But if I’m Sony, I need to worry about this, in a huge way. Do I believe that Sony replaced Amimon with SiBEAM because the former’s 1080p chip wasn’t ready, as the EE Times article states? Maybe. Of equal or greater concern is what happens when that $2000+ SKU suddenly begins throwing decompression errors due to interference on the wireless channel, one which in 2009 isn’t terribly crowded, but that in two years (only a quarter of the way into the current eight-year consumer keep of a high-end television) will be loaded like the Beltway on a snowy Friday night? That, my friends, is a risk that makes it tough for any TV manufacturer to commit to high-volume numbers.

Was Sony’s introduction of the Amimon-based solution anything but a science project? Well, yes, I believe it was. I was stoked when (with a Sony friend) I walked into the Sony employee store in San Jose earlier this year and saw the Amimon-based solution not only on display, but available for sale. However, in 2009, the high price of a transmitter-receiver pair (>$500) was never going to get Amimon to the volume needed to be wildly successful as a silicon company—a challenge SiBEAM must also overcome. Sony’s migration to 60 GHz gives SiBEAM another punch in the fight to widespread market acceptance, so I’m sure that their investors must be thrilled. SiBEAM now needs to execute successfully to reach sufficient volume that their silicon prices end up low enough to justify the added cost of a wireless solution to not just high-end TVs, but to those more accessible to the general market.

When I can walk into Costco (or Sam’s, or BJ’s) and have a choice of wireless TVs based on a common standard, only then will any company or alliance be able to claim victory.

One dark horse remains in this battle, a horse Amimon is now publicly atop. The performance of 802.11n solutions has improved dramatically over the past twelve months, to the point where I’m finally successfully streaming high definition video in my home using an off-the-shelf Netgear 802.11n access point (based on dd-wrt…w00t!). All my buddies in the connected home world made the move to 11n long ago; call me old school, or just call me a veteran of the ridiculously long 802.11n ratification process, but I’d held off on HD streaming until now. With a ratified spec in place, I’m now comfortable recommending 11n solutions to friends and family—the uptake of which raises the potential for conflict with Amimon’s existing solution with each 11n device deployed.

Now, in the truest sense of sleeping with the enemy, or changing horses mid-race if you prefer, Amimon is now attempting to align themselves with manufacturers of Wi-Fi chipsets, a move I heartily endorse. Fighting 11n was a Sisyphean battle, one that a VC-backed startup could never hope to win. Aligning with Wi-Fi vendors against the 60 GHz camp is Amimon’s best (and likely final) hope for success. Kudos for acknowledging that they need to evolve their business strategy to live, survive, and maybe thrive.

If you can’t beat ‘em, join ‘em.

One final point. Both Amimon and SiBEAM make much of their “uncompressed” video solutions. I put that in quotation marks because I still don’t buy into Amimon’s claims of uncompressed video delivery. Based on discussions I’ve had with them about their unequal error protection and Joint Source Channel Coding prioritization, I’d call their approach lossy compression, not uncompressed. If it walks like a duck, quacks like a duck, and shits like a duck, it’s probably a duck.

More importantly, you know who cares about compressed versus uncompressed? Nobody. Let me reiterate that. Nobody. Absolutely, consumers care about wired-equivalent quality, but they care more about a zero or minimal incremental cost to cut the cable. The guys with the golden eyes are still gonna figure out a way to run copper—and I know some guys who still string component, because they don’t like the video quality of HDMI, and are willing to deal with the closing of the analog hole.

Have we reached an inflection point in the wireless video wars? Yes. Is SiBEAM sitting pretty? Yes, but they have to deliver—never confuse selling with installing. Is Amimon dead? No, but Sony’s defection is a serious blow to the gut. Amimon’s bleeding, but they’re not dead yet. Getting in bed with the Wi-Fi camp has the potential to deliver a massive worldwide audience at a relatively small incremental cost. If Amimon can cost-effectively license their IP to the key players in the Wi-Fi world, adding the requisite gates to 11n silicon will be a rounding error compared to the likely cost of a SiBEAM solution over the next 12-24 months. SiBEAM’s atop the heap right now, but to stay there, they need to ship, ship, ship—and make it work.

Friday, November 13, 2009

First Impressions of the MiFi

I love me my MiFi.

I don't mean to wax rhapsodic like some central government-backed campaign, but the MiFi has brought me harmonious and glorifying wisdom, in order that my lymphatic system feels cleaner.

Or something like that.

Ever since Novatel introduced the MiFi earlier this year, I knew that it (or a derivative) would end up in my bag. If you're not familiar with the MiFi, think of a WiFi hotspot, but mobile--small enough to fit in your business card case. Seriously. The device itself is astonishingly simple--a CDMA 3G radio, a WiFi radio, and a battery. Simple, but elegant.

Over the past couple of months (or the last decade, really), I've found myself in airports, hotels, meetings, and conferences where WiFi wasn't available, or was priced annoyingly high. About two years ago, I set up Bluetooth tethering with my 2G T-Mobile BlackBerry Curve, which has three major drawbacks--2G means sloooooow; the battery drains ridiculously quickly; and T-Mobile's network is, uh, lacking. Wednesday afternoon, I'd finally had enough, so I ran to the Verizon store on University Avenue in Palo Alto to grab a MiFi. I only had about 20 minutes between meetings, so I was hopeful that I'd be able to get in and out of the store quickly--something that rarely happens when making a transaction with a mobile carrier.

Lo and behold, I got what I wanted. I walked in, told the sales the guy that I needed a MiFi with the two-year, 5 GB per month contract, and handed him my driver's license and credit card. After a double-take, he walked to the back, came out with the device, and had me out the door in less than ten minutes. THAT'S the way I like things to work.

By the way, for those of you who might ask why I went with Verizon...I seriously, seriously considered going with Sprint, since they'll be shipping a MiFi in Q1 which has both 3G and 4G support; a Verizon 4G MiFi will mean LTE, which will mean a really long wait. But, in weighing the lost productivity of the next few months without a MiFi while waiting for a 4G version, I simply said "Verizon it is". Plus, having suffered with such ridiculously poor mobile coverage (and a brutal UMA experience, which is a diatribe for another time) for the last two years, I wanted to have the best network behind me, dorky "Can you hear me now" guy and all.

One thing which had held me back is that for whatever reason, I've only seen one MiFi in the wild--a friend of a friend bought one the day it shipped, but I hadn't seen one since. I could attribute it to the fact that it's so unobtrusive, you can have it in your shirt pocket and no one will ever know you're a walking hotspot. That, or the math just didn't work for many people, but this one was easy for me to justify.

If you know me, you'll know that I'm a big out of box experience guy, and typically photograph each step of the process. Not this time--I was like a kid on Christmas day. That said, the OOBE is exceptional.
  • Open the box
  • Remove the battery door
  • Insert the battery (which was the toughest part...I see redesign in their future)
  • Replace the battery door
  • Press the power button
  • Plug it into the USB port on my Mac (I assume it works on Windows too, but hey, I don't really care)
  • Launch the disk image resident on the device itself (BRILLIANT--no CD/DVD media to mess around with)
  • Install the software
  • Activate the device using the Verizon Connection Manager (which was great--seamless self-install and activation...again, BRILLIANT)
  • Eject the drive image
  • Unplug from the USB
  • Plug into the wall (I could've just stayed plugged into USB, but I needed to charge the battery for its first time, plus I wanted to get rolling untethered on WiFi)
  • Find and attach to the SSID listed on the sticker on both the device and in the one page box insert
  • Enter the password provided on the sticker
  • Configure security (hello, WPA2; despite the WPA-TKIP I show in the screen shot, WPA2 works just fine)
  • Enjoy!
Seriously, this couldn't have been easier. From box to live in less than ten minutes, which would've taken less time if the battery installation had been a little more straightforward.

Like most folks, I was a little concerned by the 5 GB per month cap, since I really have no idea how much Internet I slurp at any given time. Thankfully, you have three mechanisms to determine your data usage: by logging into your account on the Verizon website; by logging into the device's local management screen when connected via Wi-Fi; and by clicking the "Usage" tab when physically tethered to the device via USB. I'm definitely surprised at how much traffic I'm using just by checking e-mail, but then again, I have 11 accounts in Mac Mail, and I check every minute. I'll likely kick the interval down; at 5 GB a month, that's more than 160 MB a day, but at the rate I'm going, I could envision chewing through that fairly quickly.

All in all, I'm enamored. The device occasionally disconnects, despite the fact that I've set it not to, but I simply pop into the management window and hit the connect button, and all is well in the world again. Further updates as events warrant, but for now, color me a fan. Next up: determining whether an iPod Touch + a MiFi = a Verizon iPhone. Wired did some brief testing this summer, but I hope to be able to dig more deeply soon.