Emerging Wireless Technologies: Is This "DIDO" or "FIDO"?

31 Jul 2011

The demand for wireless data services is growing by leaps and bounds while the amount of available radio spectrum remains relatively constant. The increasing stress created by exploding demand and spectrum scarcity causes us to pay particular attention to technologies that could potentially reap 10x to 1000x increases in spectrum efficiency. Such a claim is now coming from Steve Perlman, President & CEO of the Rearden Companies in the way of a new radio technology called Distributed-Input-Distributed-Output (DIDO).

The FCC is trying to free up an additional few hundred megahertz of radio spectrum, but they have to convince others (like the National Association of Broadcasters and the Federal Government) to give it up. After that they'd have to go through the political process of deciding if it will be unlicensed or auctioned, and spell out the rules for its distribution and use. If we're not all dead by then, it would still take years before they could put it into production.

In the meantime, AT&T wireless has just announced that they will begin throttling back the data rates for their highest consuming 5% unlimited data customers beginning in October. Not surprisingly, their solution is to approve their planned acquisition of T-Mobile USA.

Onto this stage steps Mr. Perlman, who while he does not appear to have any in-depth background in radio, did manage to get a writer from Bloomberg Businessweek to crown him the "Edison of Silicon Valley." While probably not on the same plateau as the Wizard of Menlo Park, in the mid-80s Perlman helped create QuickTime that delivered multimedia to the first Apple Macs. He also founded Web TV (remember that thing?), which he sold to Microsoft in 1997 for $503 million where it became the start of MSN. He holds over 100 patents with about 100 more awaiting review. He is currently President and CEO of the Rearden Companies, a business incubator he founded in San Francisco.

What caught my attention was a piece on VentureBeat.com that claims "The technology [i.e. DIDO] gets around Shannon's Law." Now in the tech world, those are fighting words. Shannon's Law, or more correctly the Shannon-Hartley theorem, is a formula developed by Claude Shannon and Ralph Hartley that provides a definitive way of determining the maximum error-free data transmission rate on any band limited channel.

Shannon published his formula in a paper in 1949, and since then it has been the equivalent of the Voice of God in transmission engineering. It has been cited endlessly to describe the maximum transmission capacity of any transmission channel (e.g. voiceband, wire pair, coaxial fiber, radio, etc.). Basically it allows you to compute the maximum capacity (in bits per second) for any channel based on the available bandwidth (i.e. the range of frequencies the channel will carry) and the received signal-to-noise ratio. Shannon's Law is something we use to determine the real impact of any development in transmission, and today the most important developments are in wireless. Over the years dozens of folks claimed to have outsmarted old Claude, but in each and every case they were proved wrong - and that's for over 60 years.

Perlman and his co-developer, Antonio Forenza, Ph.D., Principal Scientist at Rearden, have described the DIDO technology in general terms in a folksy white paper that has generated some heated commentary on the VentureBeat site. The crux of the commentary is that no one believes them or they simply see DIDO as a repackaging of other existing technologies. The best comment was from RIP1980, who wrote "The reason it's so fast is because its [sic.] lubricated with snake oil."

Unlike some of the commentators, I actually read the paper. Their basic premise is that they've come up with a way to allow multiple devices to operate on the same radio frequency, in the same area without interfering with one another. Normally if you have two stations transmitting at the same time, in the same area, on the same frequency, each one will interfere with the other. In Shannon's Law that's reflected in degraded signal-to-noise ratio, which reduces the theoretical maximum data rate.

The developers claim that they can generate a separate waveform for each individual station that coalesces around the area of that specific receiver forming a sphere or bubble they call an "area of coherence." Within that bubble, only that specific user will hear their specific signal. The receiver can also move around, the system will recalibrate, and the bubble will move with them. They don't say exactly how they do it other than an oblique reference to "involving immensely complex mathematics, very carefully designed software and hardware, and new data communications and modulation techniques." Of course, they don't build their wireless credibility blowing their description of how a Wi-Fi network operates.

What really smells like "snake oil" is the attempt to drag in as many hot terms and acronyms as humanly possible. The term "DIDO" is clearly a take-off on "MIMO" (Multiple Input-Multiple Output), the antenna technology used in 802.11n wireless LANs and 4G cellular networks. Mr. Perlman also describes it as a "cloud wireless system" - that's got to be good, right? The "cloud" reference is based on the fact that a server in a DIDO Data Center generates a waveform for each user channel, digitizes all of them, and sends them over the Internet to an access point that generates RF signals that are broadcast to all users (i.e. it's not a "directional antenna" or "phased array" trick). Each of the individual signals will coalesce or "sum" at a particular point in space so each user can hear only their own channel in their own bubble and with no interference from the others.

Now if they could actually do this (and I have no reason to believe they can based on what they've said or demonstrated) it would be one heck of a trick. It still wouldn't negate Shannon's Law however. The capacity of each of those coalesced channels would still be governed by bandwidth and signal-to-noise ratio parameters Shannon defined. If they could effectively cancel out all the interference from the other channels, the signal-to-noise ratio would improve, though it wouldn't have any impact on RF energy from other sources in the same frequency band.

If DIDO turns out to be real, it could be a really big story (though a 1000x performance boost is hard to buy). In any case, it wouldn't be the first such story we've seen in wireless. Technologies like code division multiple access (CDMA) and MIMO are pretty far past what they taught you in high school physics; many people doubted those advances at the outset. So Messrs. Perlman and Forenza have thrown down the gauntlet. If DIDO is real they could make them a fortune. On the other hand, if DIDO turns out to be "FIDO" ("Flaky Information for Deliberate Obfuscation"), the VC's will tell them to "go bark at the moon."