Thursday, March 12, 2009

Attack Vector Risk Management

I read this post from Michael Dahn and I really liked what he called "Attack Vector Risk Management". Today I saw that the guys from Sensepost also noted the post for the same reasons, and even showed some of their work under the same concept, calling it "Corporate Threat Modeling".

During the last months my main interest is enterprise security planning. How should an organization define how to spend its security resources, what should be done and in what order? Risk Management is usually the answer for that (please DON'T SAY COMPLIANCE!), but IMHO the risk assessment methodologies out there just don't scale to a point where they can be used to drive security decisions in an enterprise level. You start using so many "educated guesses" that the end result is just not intellectually honest, everything is extremely biased to what people believe that are their major risks that just a simple brainstorm would probably generate the same results. Have you ever seen the results of an enterprise level RA being a surprise to anyone (except for dumb as hell CISOs!)? I haven't.

I don't think that Sensepost approach escalates well too, but it seems better than regular RA for me. I believe we can come tom something that is "threat oriented" than can generate a better understanding of an organization security requirements and help the development of a security strategy. After that we will finally be able to bury ROI/ROSI stuff and stop pretending that those beautiful tables with numbers, "high/medium/low"s  or "green/yellow/red"s are something more our minds tricking us into believing that there is a mathematical explanation behind our intuitive perception.

Until there, you can read "Blink", from Malcolm Gladwell (yes, the guy from the current best seller, "Outliers"), to see that simply trusting our intuitive side is not that bad, although I just can't see a CISO telling an auditor that his security strategy is "intuition based" :-)


No comments:

Post a Comment