The Lectern: Cyber-security -- protection vs. resilience

I had dinner recently with Dan Mintz, the Department of Transportation's delightfully smart and contrarian chief information officer, and asked him about what his priorities were going to be for the waning months of the administration. Security, he said. Sort of a conventional response. But Dan's not a conventional guy, and he proceeded to develop his frankly unconventional views on cyber-security priorities.

We need to avoid putting too large a percentage of our cyber-security resources into firewalls and other forms of protection against people breaching the fortress and getting access to the system, in Mintz's view. No matter how hard we try, we will not be completely successful, and the concentration on defending the fortress acts psychologically to create a Maginot Line mentality where we don't think enough about what we should be doing behind these firewalls.

He argues that we should be putting a greater proportion of resources into what he called "resilience," increasing our ability to survive successful breaches; What does he include under this rubric? Partly, getting better at detecting breeches, to give us more time and ability to react. Partly, getting data kept in more dispersed locations, so crucial parts of our systems are more backed up. And partly, architecting our systems to be more resilient to attack, which broadly speaking involves getting them more Internet-based architecturally, so they become more decentralized and thus harder to bring down.