March 18, 2017

"In general, there are two basic paradigms of security. We can either try to secure something well the first time, or we can make our security agile."

New York Magazine: The first paradigm comes from the world of dangerous things: from planes, medical devices, buildings. It’s the paradigm that gives us secure design and secure engineering, security testing and certifications, professional licensing, detailed preplanning and complex government approvals, and long times-to-market. It’s security for a world where getting it right is paramount because getting it wrong means people dying. By Bruce Schneier

'The second paradigm comes from the fast-moving and heretofore largely benign world of software. In this paradigm, we have rapid prototyping, on-the-fly updates, and continual improvement. In this paradigm, new vulnerabilities are discovered all the time and security disasters regularly happen. Here, we stress survivability, recoverability, mitigation, adaptability, and muddling through. This is security for a world where getting it wrong is okay, as long as you can respond fast enough.

'These two worlds are colliding. They’re colliding in our cars — literally — in our medical devices, our building control systems, our traffic control systems, and our voting machines. And although these paradigms are wildly different and largely incompatible, we need to figure out how to make them work together.

'So far, we haven’t done very well. We still largely rely on the first paradigm for the dangerous computers in cars, airplanes, and medical devices. As a result, there are medical systems that can’t have security patches installed because that would invalidate their government approval. In 2015, Chrysler recalled 1.4 million cars to fix a software vulnerability. In September 2016, Tesla remotely sent a security patch to all of its Model S cars overnight. Tesla sure sounds like it’s doing things right, but what vulnerabilities does this remote patch feature open up?'

No comments: