Greg Ness: Key to Virtualization kingdom is security!

Just like Applications , as I mentioned in the previous post, security will also change dramatically.

Changes can be Painful for many Security Solutions

In production environments effortless movement and changes of VM states (snapshot, revert, online, offline, VMotion, etc) can generate extreme operational challenges for critical security activities like vulnerability scanning, patching and security. Vulnerability scans, a critical tool for tracking software vulnerabilities, can become obsolete in seconds. Bottom line: The constant change enabled by virtualization can place dynamic demands on the most commonly deployed static security solutions, in even small virtualized production infrastructures.

All the tricks, flips and tools that make software more nimble and powerful will not matter unless the production infrastructure can be effectively secured from attack. Yet many of the leading network security vendors have been caught flat-footed by virtualization. Some are even trying to cram ASIC-driven IPS solutions onto commodity processors, taking up sizable chunks of server/blade processing power and introducing unacceptable levels of latency, in a nonsensical effort to match suspicious virtual server traffic with a growing library of signatures. That game promises to get even more complicated and resource-consuming as hackers shift to mutating attacks.

Taking the challenges a step further into the virtualized data center made up of blade server fabrics: just how many enterprises will be returning to the ASIC security world with bigger boxes, bigger signature libraries and the promise of constant tuning and traffic headaches while the rest of their infrastructure becomes more powerful, more flexible and more efficient?

How many ASIC-driven security players (and their hardware-centric channel partners) are likewise talking a hard look at the pure software model of virtualization (and much lower margins) and seriously contemplating “serving up their children and their channel allies” to deliver a core technology that in its current state is likely unfit for commoditized processing? That’s an Innovators Dilemma that might even make Clay Christensen cringe.

Clay's my hero too, Greg ;-)

Strongly suggest you to read Greg's post.