What to Watch (& Avoid) At RSAC

Endpoint Protection, including, for example, “next-gen,” money-backed guarantees, and other “gimmicks.” It seems to me that these technologies are always trying to keep up versus actually being proactive by constantly learning and adapting to new threats and attacks. 

SIEM. This is perhaps one of the least sexy sectors in the vast cybersecurity landscape. As more and more devices become part of the overall elastic network perimeter, the rate of data being produced is explosive. Once again, this is a technology that is always trying to catch up even though the operational costs of running and maintaining it are not trivial.

Incident Response Automation. (Otherwise known as, “How can we get the post-mortem completed as quickly as possible?”) Instead of trying to solve for something that’s already occurred, how about we start focusing on actually preventing the majority of incidents from happening in the first place? Advancing automation for more secure products and processes is good but right now the emphasis is in the wrong place.  

IoT Security. Internet-enabled now seems to be table stakes for any new device being released to the market. Given the recent large scale attacks caused by IoT devices, there needs to be an overall standard, and I really hope that UL starts being more transparent. Its Cybersecurity Assurance Program is a good start, but there needs to be far greater diligence applied to it.

Identity Management. Infrastructure has transformed into software-defined and elastic, yet most identity provider solutions still remain quite rigid. Many also only focus on the AuthN (authentication) part of identity versus AuthZ (Authorization),  where innovation should occur, combined with simple best practices such as “least privileges.’’ In the age where everything has an identity, there should be a platform that is adaptive enough to support this. 

Artificial Intelligence + Machine Learning. Skills shortage or not, scaling out human capital to attempt to keep pace with "The Singularity" is not an option. Continuous analysis and learning using a variety of techniques, including Behavioral Analysis and Game Theory, is what is required to truly move the needle in cybersecurity. This area also overlaps with my next area - DevOps - and the goal of creating a true culture of DevSecOps. In other words, an AI/ML solution has to be an API-driven platform solution, not another point-solution tool. (Note: If you are learning or hyperfocused on AI in general, you may want to check out Gigaom AI, also taking place in SF during RSAC).

DevSecOps. One of my core assertions is that security engineers need to adopt a software engineering mindset and approach to solutions. This is not dissimilar to what happened to “classic” system administrators during the shift to DevOps. Security needs to be seamlessly integrated into the entire software development lifecycle, instead of being a barrier to deploy and the “Department of No.”

10 Essential Elements For Your Incident-Response Plan

The Promise & Peril Of The App Era

Cloud Security & IoT: A Look At What Lies Ahead