A lot of people have asked me over the last few months, "What is the difference between Pervasive Computing and Wireless Sensor Networks." The quick answer is that WSN nodes are traditionally "dumb" -- they just sense data and forward it to a server for analysis. In contrast, pervasive computers are exactly that -- computers -- which means they are intelligent and can make their own decisions.

A common analogy is the era of mainframe computing. Users commonly worked with "dumb terminals" connected to a mainframe. The terminals were simply remote I/O devices (keyboards and screens) with all the intelligence in a centralized point (the mainframe). But people started to realize that they could take the terminals home with them, share data (like spreadsheets) with others directly, and thus the PC revolution was born. Similarly, WSNs today are "dumb terminals" that have a wireless tether to a mainframe (often called a server or gateway). In contrast, pervasive computers are the next generation technology -- they are the PCs that can intelligently run applications. And similar to PCs, there's the need for a software platform to build millions of killer applications.

To explain more about what this all means for developers, IT managers, and our peripherals (like sensors and actuators), I wrote a contributed article for Sensors Magazine that was published this week.

Read the article for more information and post your comments to Sentilla's blog.

Update 11/20/2007: An extended version of "A New Vision for Pervasive Computing" that includes additional applications and some additional analysis of available solutions is available here .