Your Car: The Next Target of a Cyberattack?
There are so many new security risks emerging all the time that we barely have time to keep up with any single one. But there is a new, potentially horrific risk that demands our attention—“car hacking.” Researchers at the University of Washington and University of California-San Diego have discovered a way to attack a network that is built into most automobiles nowadays. This network, called the “Controller Area Network (CAN),” provides control over numerous functions such as anti-lock brakes and anti-rollover mechanisms. It also regulates fuel flow to save gasoline. The researchers showed that in cars with built-in wireless networks such as Bluetooth it is possible to gain unauthorized remote access and then inject code that initiates undesirable and potentially unsafe actions such as making a car unable to brake even though the brake pedal is pressed, turn the engine off, sound the horn, lock the doors, open windows, turn the air conditioning and the radio on and off, display falsified speedometer readings and more. The researchers have demonstrated proof-of-concept at an abandoned airport in the state of Washington by running a program called “CarShark,” which remotely sends control instructions to CANs with no authentication whatsoever required.
The good news is that the likelihood of successful “car hacking” in real-life settings is low. A car hacker would have to write a very sophisticated program and would have to be physically near a targeted car. Still, sophisticated programs such as CarShark always seem to fall into the hands of amateurs sooner or later. The fact that no patch or other method of warding off “car hacks” is currently available is genuinely troubling, but the fact that the black hat community is likely to turn much of its attention to “car hacking” is even more disconcerting.
How is it possible that all these malicious and unauthorized actions can be initiated by someone who is up to no good? The answer lies in ANSI standards ISO 11519 and 11898 for CANs; both are functionally void of security-related requirements. And apparently the fact that automobiles’ CANs could be remotely attacked never even remotely dawned upon car manufacturers.
Once again we see a case in which security has been overlooked in international standards. We all know that security must be built into something right from the start if security is to work optimally and be cost-effective. Retrofitting security is better than nothing, but retrofitted security is often easier to defeat than security that is built-in up front, and it is also generally more expensive. Apparently, ISO is either not aware of these critical principles or too many “security naysayers” were on the teams that wrote the ISO 11519 and 11519 standards.
A final thought—you undoubtedly know about all the trouble some models of Toyotas have had with unexplained acceleration and brakes not working. Hmm, imagine these kinds of problems in most recent models of cars due to “car hacking.” In a way, Toyota might experience a sense of relief if “car hacking”