Following the news that Mitsubishi Outlander hybrid car alarm ‘hacked’, here below the comments from two different sources.

  • Richard Kirk, SVP at AlienVault
  • John Smith Principal Solution Architect at Veracode

Richard Kirk, SVP at AlienVault:

Richard KirkMitsubishi have taken a different approach to enabling access to the Mitsubishi Outlander hybrid. Whereas most, if not all other car manufacturers provide support for their car apps through a cloud based service, which in turn sends instructions directly to the car, Mitsubishi appear to be using an onboard wifi network which the app directly connects to. This might appear to offer support for cars that do not have a mobile data connection (for example, half way up a mountain), since the app would be communicating directly with the car via the onboard wifi system. However this opens up many potential security issues, which have since been discovered. One has to ask why the app developers did not fully explore all the potential attack vectors, including a visible wifi access point, which is like leaving a back gate open. Users should disable the service until Mitsubishi provides a fix, since there isn’t a way to override how the onboard wifi access point operates. Mitsubishi, like all car manufacturers, should involve their security teams during the design and development of their in car app services, and perhaps consider employing a security monitoring managed service to be able to detect unusual behaviour.”

John Smith Principal Solution Architect at Veracode:

John SmithThis bug is the latest in a growing line of connected car vulnerabilities that puts the safety and security of the car – and its owners and passengers – at risk.

Recent Veracode research, conducted by IDC, found that there could be a three year lag before in-car technology and automotive applications are developed with security in mind. With this in mind, it is critical that car manufacturers learn from these examples, ensuring that all driving applications are developed with robust cybersecurity methods from the outset.

1 Expert Comment
Newest
Oldest Most Voted
Inline Feedbacks
View all comments