Rather, he pointed out, many of the challenges associated with virtualization are operational. Such as taking an unsecured virtual machine that's not being properly patched and using that VM as a template that is then spread about your environment. Then those VMs are cloned, and those unsecured machines spread quickly.
The other organizational issue is the collapsing of multiple IT responsibilities -- such as security, storage, access rights -- to a single virtualization admin, thereby removing the separation of duties many companies rely upon.
The third and final operational virtualization security challenge Perilli covered is the consolidation of different risk levels into a single machine. For instance, a company may have a server hosting a VM with highly classified information alongside several VMs managing what is considered public information. Not good for managing risks on all of the VMs in that physical host, or virtualized zone. Then there are implementation issues, which are caused in large part by the lack of mature native VM management capabilities: think weak access controls for multiple VMs accessing a shared storage pool. Then there is the fact that many security settings and other controls don't properly follow VMs when they're dynamically provisioned.
Let's not forget the fact that software itself, today, is inherently unsecure as it's developed. "Any software can be compromised," Perilli said. "The same will be true whether it's ESX or Hyper-V."
The problem of attacks on virtualization software will only become even more pronounced as vendors, as has been the case with operating systems, cram ever more features and functionality into the hypervisor. In fact, according to some stats Perilli provided, it's already happening: in 2006, VMware published less than 10 patches. That figure climbed to nearly 50 patches in 2007, and more than 60 patches being published already this year.