Interview You want cheap Wi-Fi? "You'll get what you pay for," warns Selina Lo. "Wi-Fi makers are in denial. The system doesn't work properly, and will get worse."
Thank God, she says happily, "for manufacturers like us!"
No, overweening modesty isn't her major personality flaw. And as CEO of Ruckus Wireless, her third start-up company (previous success was Alteon), it's not obvious why she should suffer from it. Anyway, I agree with her:
Ruckus president and CEO, Selina Lo.
"Wi-Fi is a selfish technology. What people need is the ability to make Wi-Fi behave. That's what we do - we make Wifi into a utility."
For anybody who lives in a congested wireless neighbourhood, the "will get worse" prediction is already coming true. You go to the computer store, you buy your Wi-Fi access point, and then you find that you aren't the first; and you also find that it's slow. Promises of "faster than wired Ethernet" are made - promises which don't translate well into reality.
If you think your Wi-Fi is the exception, Selina Lo has an acid test for you; try running IPTV across your wireless link. With a Ruckus wireless link, you can.
This, says Lo, explains why all the "triple play" service providers are queueing up to buy Ruckus technology. It's a technology with the snappy name of MediaFlex: "The first in-home multimedia 802.11a/b/g system developed specifically for service providers" says the blurb.
Sitting listening to Lo explain all this, I had to admit that she doesn't present the typical image of a CEO of a high-tech start-up. She is (you should excuse my directness) tiny and cute. If you're looking around the hotel lounge for her, and spot her, you might easily notice her, and pass her by as being far too young anyway, and lacking in the sort of charisma which the typical American VP tries to cultivate.
Listen to her talk, however, and the charisma hits you between the eyes.
The technology that Ruckus is hoping to make into an industry standard can't be done, Lo insists, with "the standard cookie-cutter output from the standard one-chip solution. You have to do some actual R&D - you know, wireless research."
Isn't that what all the wireless companies do?
"No. Nobody does any R&D on Wi-Fi layer. Big companies, aiming at the Fortune 500 market - people like Cisco and Aruba - do research into management and other upper layers of security - not into making the RF layer robust. And that's true for the chip guys, too - from their perspective they want to maximise the market for the same chip. They want to produce something that will work in all situations, from consumer to enterprise."
And of course, Ruckus reckons it has done the research that matters: into the middleware of the wireless layer. "This isn't software middleware. This is wireless middleware."
The jargon, when analysed, boils down to a smart antenna system; a combination of very tightly designed antenna that is configurable by software: "And we provide the software which drives it. Each antenna system has multiple elements - the software configures which elements to turn on - on a per packet basis."
So the MAC layer is designed to ACK every packet sent.
"We found that by monitoring the antenna system, we learned about every wireless destination - when you get a poke from someone, you get a lot of information and we use that information to build up a database. So we have signal paths formed by different antenna elements to different destinations. We can focus on the correct signal, even if there is noise being picked up by the other antennas."
Radio frequency interference, or noise, is a black art. But it's not beyond understanding. Anybody who has carried a portable radio and discovered that if you twist it to one particular angle, it loses the signal, is demonstrating the Ruckus effect. Find out which antenna elements are at the right angle to ignore the RFI, and you can screen out the noise.
You don't meet many CEOs like Selina. She tells me her own expertise, way back, is as a Comp Sci graduate, not wireless - but she got into HP in its glory days (before the engineering half was split off) and found herself actually running things, rather than just tinkering with code and systems.
When she exited Alteon, she decided to take a holiday. It was a long one, ending up eating much ice cream in Italy... and trying to install TV in two rooms in her home in Silicon Valley. She wanted it in two rooms; to do that, she found herself spending thousands on rewiring and ended up with a socking great hole in the ceiling, watching the installer trotting off complacently explaining: "We don't do interior work."
"So I thought, why can't I use Wi-Fi for the TV signal?" She found a couple of guys trying to start a company up which could. And they couldn't get funding "because everybody said Wi-Fi is a commodity, so the three of us put up a million and a half dollars of seed capital, and the VCs like Sequoia followed."
Lo, it turns out, was born in Hong Kong, and originally went to America with her family, aiming to learn enough English to get a degree which would enable her to teach. While she was studying, things in Hong Kong changed and she felt less and less like teaching English in the colony, so switched to UC Berkeley and computers.
Now, with MediaFlex starting to sign up big carriers like Telefonica and PCCW and a soon-to-be-announced European contract with one of the big five, Ruckus is starting to promote its latest baby: ZoneFlex.
"It's aimed at the Unfortunate Five Thousand," summarises Lo. "The middle size company can't run to the big-bucks networks with all the bells and whistles, but they can't manage by buying the cheap stuff, because it won't scale. We can handle their needs. We recently deployed it in a Holiday Inn, a building with 100 rooms, 3 stories, 18,000 square feet per storey.
"It had only one Ethernet port per floor. We deployed Zoneflex in this Holiday Inn with less than a dozen access points. One on each floor has the AP connected to Ethernet - ll others are wirelessly connected through that one. It took us three hours, 10,000 dollars to cover the hotel."
She's preaching to the choir, with me. In the last year, I've watched my inner London network slow down further and further, until it really isn't something that you can rely on. The 802.11n MIMO technology will make things worse, and the collapse of the ultra-wide band standards process means that there's no hope from that quarter, even with the Bluetooth SIG working on the problem.
Sadly, the one thing they can't do ("but we're working on it!") is synchronise the stream from one TV antenna, to two sets. Digital streaming doesn't work like that; there is a variable latency much of the time, and so when you walk from kitchen to living room, there will be a chance that the one set will sound like the echo of the other.
"We can live with that. The important thing is that the carriers can install IPTV in someone's home quickly, and without pulling co-ax cable through holes in the walls. And if we can solve the IPTV problem, which is the hardest, we can solve the others too." ®