updated 11:10 am EDT, Thu July 28, 2011
DIDO technology promises to get 10X capacity
Acting on promises, OnLive founder Steve Perlman on Thursday published a white paper (below) outlining a technology that could theoretically clear up cellular and other Wi-Fi networks. Distributed Input, Distributed Output (DIDO) skips computing the direct signal waveforms of most technology and instead has a datacenter handle the task. Because the datacenter can produce different waveforms, each device gets its own distinct waveform and doesn't face interference, even if a more powerful other signal is in the area.
The process would circumvent Shannon's Law, a rule of wireless that dictates a degradation of speed as more users share the same frequency and channel. DIDO works by effectively creating a new channel for each person on the spot.
Signals also work up to very long ranges of up to 250 miles, since they can bounce off of the ionosphere and don't deal with the basic curvature of Earth as a range limit. As an example, Perlman's team set up access points in Austin and distant neighboring towns that could all connect to the same DIDO site. Latency is also much lower at under a millisecond in usual conditions and no more than two to three in a rural layout.
If implemented, DIDO could increase the capacity on a given frequency by at least 10 times, with 100 times considered possible, Perlman said. He was also "optimistic" that a 1,000-fold increase was possible. Research still hadn't shown slowdowns, however, and made it difficult to tell where the limit would ultimately be.
The OnLive head also stressed that DIDO was meant to be a real, commercial product that could be used. The design was not only intended to scale but could actually be cheaper and more efficient than existing hardware in networking devices. It needs less power to get the same range and can work on radios simpler than cellular and Wi-Fi.
Perlman didn't give signs of how soon he expected DIDO to translate to real products. Using the technique would involve a substantial change to wireless as a whole, though, includning both access points as well as individual devices. The invention could improve on both cellular and Wi-Fi, however. One possibility is rural wireless broadband that would be cheap to deploy but wouldn't bog down quickly with more users. It could theoretically end much of the "spectrum crunch" that has inspired deals like AT&T's buyout of T-Mobile by letting companies handle many more users with frequencies they already use.
Perlman's lecture on DIDO from June; starts in earnest at 28 minutes