Yes, it’s happened. Radical new wireless technology could increase data speeds by a factor of 10. Maybe 100 or 1000. Shoosh, Mr Abbott, it won’t kill the NBN. But it could transform the world.
Steve Perlmann has just released a white paper describing DIDO, “distributed input distributed output” wireless technology.
What he’s suggesting should be impossible. On the surface DIDO appears to break Shannon’s Law, a fundamental theorem of communications that has shaped our views since the 1940s.
Shannon’s Law underpins the fact that all the users on a wireless base station have to share the capacity. The more users, the lower the data rates they can each achieve. Interference from neighbouring base stations or other transmitters reduces their data rates even more.
Think of each base station being a room full of people. The more people talking, the harder it is to hear what’s going on. Communication slows down as people shout to be heard and repeat what was mis-heard.
There’s a longer explanation of how this applies to wireless data at VentureBeat.
Perlmann’s idea is revolutionary. DIDO takes constructive advantage of that radio interference.
Each user listens to all the wireless base stations in range. Each station is transmitting a unique, complex signal that’s been precomputed precisely so that when they arrive at the user’s location they interfere with each other to magically produce the clean signal you wanted to deliver there.
Every user gets a data rate equivalent to the full Shannon’s Law limit as if they’re the only user.
“How is DIDO able to create waveforms that, when they sum together at each user results a clean waveform with the data for that user?” write Perlmann and his principal scientist Antonio Forenza.
“The complete answer to this question is very long, involving immensely complex mathematics, very carefully designed software and hardware and new data communications and modulation techniques.”
OK, we’ll skip ahead a bit.
None of this would be possible without massive computing capacity. But we have that now.
If you’re one of the 26 people in Australia who understands how holograms works, well, I reckon it’s kinda like an inside-out hologram. Or something. I’m also thinking phased-array radars and things like Australia’s Jindalee Operational Radar Network.
My. Mind. Is. Blown.
If this was anyone but Perlmann, I’d file DIDO over with the perpetual motion machines.
But Perlmann has cred. He invented Apple’s Quicktime video technology. Business Week calls him the Edison of Silicon Valley. He founded WebTV and sold it to Microsoft for half a billion dollars. He’s … a genius.
Perlmann and Forenza say they’ve got DIDO working at test sites in Texas. “We’ve tested DIDO at frequencies from 1 MHz to 1 GHz, and it works great at all frequencies.”
Now I can already hear the NBN’s opponents bragging that they were right, that new wireless technology would make fixed fibre connections obsolete. It’s true that if Perlmann is right, the limits we’d assumed for wireless have just been be blown away. But announcing fibre’s death is a tad premature.
We’re building the NBN today. To repeat what should be obvious, we can only build it using technology that’s available today from vendors who can deliver millions of units and have training and spare parts and technical support and all those supply chain issues sorted.
DIDO is only still only an experiment. Three users on the plains of Texas is well short of 20,000 users in the Sydney CBD, where signals are distorted by buildings and half the users are moving around. Scaling up DIDO will be a significant engineering feat. It’ll be years before DIDO is available commercially.
It’s not like do refuse to buy a car when we know there’s be a newer, better model next year. And it’s not the like fibre will suddenly stop working. Besides, Perlmann may still be derailed.
“Wireless behaviours unique to DIDO — including the most intriguing ones — are still being studied,” Perlmann and Forenza write. “We have observed wireless phenomena that we do not believe have been seen before that will take time to understand and document. Stay tuned …”
Wow. Just, wow.
Interesting, but Perlmann’s work needs to be reviewed by the people who understand the science before his findings should be accepted. Then engineers need to examine whether Perlmann’s proposal is practicable and business analysts are needed to cost a project.
@Gavin Moodie: Agreed, there’s probably a long way to go yet.
And I have another question. What is the upstream data rate? Everything I’ve read talks about the downloads. Great for 2000 people watching 2000 different video streams. What if you want to send 2000 video streams?
Since we’re talking about barely past the conception stage communications technology, how is the performance of the newest experimental fibre-optic transmissions ? Less or more than 1000x faster than this “DIDO” thing ?
It sounds like Perlmann is claiming to have reinvented http://en.wikipedia.org/wiki/Multi-user_MIMO
Or is there something different that I’ve missed?
More to the point, how many users did they test DIDO with? I can imagine the apparatus being adjusted for optimum performance for 1, 2 or 3 users. What about thousands of users?