The Ultimate Limitation

We've made great progress, across the board, in wireless - higher throughput, lower prices, improved security and manageability, and more. But I have to say that I'm not really all that surprised that, for example, that we have 100+ Mbps WLANs. The reasons I got into wireless in 1991 were (a) the simple notion that mobile people with mobile computers need mobile (and that means wireless) networks, and (b) that VLSI would be just as effective in wireless as it had been (and continues to be) in microprocessors, memories, and other computer components. I expected we'd also make great progress in protocols, channel coding, and, well, everything, once item (a) above became common knowledge. And it did, and here we are, and we're not done yet.

In fact, there is a fundamental limitation, especially as we add ever more functionality and throughput, and that's power. Note a discussion of power has two elements, the demand side and the supply side. In terms of supply, that's traditionally meant batteries. We've seen good progress in batteries, and innovation continues here - I'll have more on the supply side tomorrow, I think. The bigger challenge has been on the demand side - more functionality, higher clock rates, and power demands just skyrocket. Sure, we've got power-saving modes on mobile devices, and power-conserving protocols over the air, but the capacity of batteries themselves must be limited for reasons of mobility (size and weight) and, of course, cost.

And there are some applications where power is really critical because battery life must be measured in years. I spoke yesterday with Cees Links, who's the CEO of GreenPeak. You may remember Cees (pronounced "case") as one of the founders of the 802.11 effort back in 1991, when he was working at NCR on WaveLAN, one of the first wireless LANs. Today, though, GreenPeak is building ultra-low-power mesh radio nodes for sensor applications. Cees mentioned to me a story about a sensor mesh with 4,000 nodes and each node having a battery life of 10 years. Sounds good until you realize that you'll still have to change a battery on average every day!

GreenPeak's products (announced today) are compatible with the ZigBee spec, but their energy consumption is so low they can be powered via energy harvesting. The idea here is pretty simple - capture energy from light or motion (the latter using piezo actuators or similar) or other source, and perhaps never change a battery. This is cool, and may point the way to longer battery life in higher-order electronics like PDAs and other handsets. Think self-winding watch. Remember those?