Perfect Parking ... Defeated?

Finding a parking space may be easier in the near future, thanks to mobile apps that can pinpoint available spots on a map. However, in the rush to install such technology, technical issues remain unresolved.

Using a $19.8 million grant from the U.S. Department of Transportation, the SFPark program uses thousands of sensors dug into the concret throughout the city of San Francisco to detect available parking spaces. The system also taps into the city's new smart parking meters. Based on use, the city can change what it charges for each of the open spots.

An even larger effort is underway with Streetline, a company that bills itself as "reinventing parking." There are two models of ParkSight: one for municipal governments, and one for private companies such as airports and garages. Central is the company's package, which includes surface sensors and the back-end analytic tools.

Both SFPark and Streetline's ParkSight rely upon mesh networks, a series of repeaters that bounce the sensor signals back to a few data centers. Such networks can support between 100 to 5,000 nodes.

There are technical challenges, such as electromagnetic interference from buried power cables, or sensors that have been blocked by obstacles, such as garbage containers. And then there's weather or construction, both which can obscure or even dislodge the embedded sensors.

What's not clear is how secure these signals are. Neither SFPark nor Streetline discuss encryption on their sites. Which is a problem.

In the summer of 2009, Joe Grand, Jacob Appelbaum, and Chris Tarnovsky showed at Black Hat USA how they could spoof a legitimate payment to one of San Francisco's new smart meters. While spoofing parking availability might not sound so bad in comparison, it defeats the whole purpose in having the parking sensors out there in the first place.