"We see a sea change coming in microsensor technology," says Stonebraker, the recent 2005 recipient of the IEEE's John von Neumann Medal. Microsensors will be used not only to track pallets of goods but, eventually, each item on the pallet.
Tags also could be used in more pedestrian ways. An RFID-style tag on your car will identify the vehicle to the parking meter when the time on the meter expires, which will then broadcast a report that guarantees you will get a ticket rather than trusting to a hit-or-miss encounter with a meter maid.
"Microsensor technology is going to solve a lot of social problems that have been hard to address," says Stonebraker, although he's thinking on a somewhat larger scale than parking meters.
Is traffic backed up at a traffic choke point, like the Bay Bridge on Route 80 in San Francisco? In the future, the bridge fare may be based on how many people are trying to get across the bridge at a given time, with fares steeper during peak periods. The current $5 fare might zoom to $15 during peak congestion and drop to 50 cents during low traffic, explicitly encouraging those who don't have to be there during rush hour to wait for another crossing time. Such fares would be based on reading the RFID tags on cars as they pass a point approaching the bridge, with the concentration of cars becoming a predictor of bridge traffic.
The arm band on a child at Disney World will have an RFID tag that tells a parent where the child has strayed to if they become separated, he predicts.
The RFID tag on your laptop will communicate with the alarm system as it's carried out of the building. If microsensor messages on both carrier and computer don't match, an alarm will bring security staff running to see who's making off with someone else's computer, he suggests.
"We see a giant market to deal with this fire hose of sensor data," Stonebraker says. He is a founder of StreamBase, which produces software to analyze streams of data in real time, a task that's difficult to do with relational databases because they lack a time window, or set period in which they are watching for incidences of particular data.
He gives an example of weather sensors producing data that indicate high winds from a storm are about to strike a ferry in Baltimore harbor, an incident that actually occurred, capsizing the ferry. By analyzing the weather data from sensors afterward, both the force and path of the wind can be plotted. To be able to do so in real time on a stream of weather data would allow a warning to be issued to the ferry captain, he says.
Right now, the main users of the StreamBase product are on Wall Street, analyzing streams of stock-ticker data that can help decide whether to buy or sell certain issues. But once RFID and other sensors are everywhere, the task will become much larger. Analyzing streams of data and drawing conclusions from the stream -- this wind is going to capsize this ferry -- will be a powerful tool in the hands of many businesses, he says.