It's been said over and over again that there's no difference between the administration of a physical server and a virtual machine - and this is true in an absolute sense. However, this might give us the idea that a virtualised datacentre is administered in the same way as a physical environment; thinking like this leads us to the first pitfall of virtualisation. Especially when dealing with large-scale virtualisation projects, we need to rethink network uses (workflows, operating procedures) and reposition IT professionals so that they can provide the best technical expertise in these new technologies.
It's true that virtualisation facilitates the repeated use of the infrastructure. The most striking example is the ability to set up several virtual machines with only one template. In just a few minutes and only 5 mouse clicks, a new (virtual) server that's fully operational and in line with company standards can be set up.
The extreme simplicity of using this functionality leads to the dangers of the virtual environment: saturation. It's so easy that anyone can do it... even without thinking... and the risk is that this creates major problems that can appear 6 months later when the infrastructure has been saturated with too many machines
So how can you avoid this? From the beginning, before setting up any virtual infrastructure, you need to study operating procedures and adapt them to the new infrastructure. In this example, one solution is to introduce a workflow for requesting the creation of virtual machines with an approval process that helps manage and commission resources. Similarly, it may prove useful to apply decommissioning dates (renewable) for test machines in order to limit "pollution".
In any case, we can say that for a new virtual infrastructure to offer its full potential, processes must be adapted to the new tools. This also applies to IT professionals.
Because it's so centralised, virtualisation tends to dichotomise roles within information systems hierarchies. Virtual environment administrators end up working on the system (VMs), network (Vswitchs), storage space (Datastore) and server resources. The problem is that network and storage administrators are unaware of what goes on in the virtual environment, since they don't have visibility on it. Due to the division and separation of teams, a lot of IS maintenance and updating operations become counterproductive or cause negative effects that freeze the system. Haven't we all seen network or security teams stop work on projects already well underway because they needed more information (communication) about the system's intended organization?
Especially when setting up a virtual infrastructure, it's best to rethink the proper role and place of each infrastructure professional. We need to imagine new, transversal organisations in which IT personnel can work in a different way.
In my view, virtualisation allows administrators, engineers, and architects to recentre their work on its true value: engineering and design. The installation of a virtualised infrastructure leads to quantifiable benefits in recurring administration tasks (installations, updates) through the industrialisation of processes. This means IT personnel once again have the chance to focus on efforts that will earn new profits for their company (streamlining flux processes, offering business development tools, improving existing systems, etc.).
We can see that a virtualisation project is most valuable when we know how to go beyond mere technical installations and learn to rethink the uses of a network and the roles of IT professionals within an information systems department.
So, to ensure a project's full and total success, the first action to undertake, before proceeding with technical integration, is to study how the IS is administrated and plan the changes to be made in the network uses and the roles of IT professionals.
I work around the concepts related to the rationalization of infrastructure information systems and particularly on user environments transformation projects.