Security recommendations change depending on whether they deal with a server or an appliance. How does that mesh with a virtual infrastructure in which both ESX and ESXi play almost identical roles and have similar vulnerabilities?