As I mentioned last week people have got used to the idea of having their own windows systems for their own section of work and have been slow in adopting a more resource sharing attitude - this means that each server is doing very little.
Has the move to virtualization technology corrected this issue?
Virtualization allows multiple Windows systems to run on one physical machine as ‘guests’ under a hypervisor and this in theory should make the physical system utilize more resources but the problem lies in the mentality of the people running the system not the technology. Packing guests together means physical machine utilization should be higher, but a number of problems still persist:
- “MINE!” still prevalent - Being used to having a one server per service environment staff create the same situation with Virtual machines, but now it has been made even easier to have multiple machines and often they’re not even very busy!
- Easy to suffer virtualization sprawl – this is where the number of virtual machines (VMs) on a network reaches a point where the administrator can no longer manage them effectively
- Virtualization sprawl is often for “good reasons” – such as building redundancy into the system, this means if one VM is taken down another one can be put up with virtually no downtime.
- Some organizations rely on high-availability / dynamic resource sharing – Careful Planning is needed to make sure that components of a service do not end up together on the same physical machine as if it fails it takes down a whole service.
So how do we properly capacity manage a Windows environment? I'll be talking about this on Wednesday.
In the meantime sign up for our next webinar 'Capacity Management from the ground-up - a case study'