Concepts
When discussing the challenges that server virtualization imposes on capacity management and
IT Service Optimization, we need a common understanding of the mechanisms that comprise
a flexible virtualized environment. Conceptually, these are the three mechanisms used:
1. Virtualization of Resources grouping resources into pools that are assigned to different
virtual machines by a hypervisor.
2. Dynamic Resource Scheduling the ability of the hypervisor to automatically assign and
direct resources to virtual machines based on their current demand.
3. Seamless Migration the ability to do resource scheduling across separate physical
servers, enabling the implementation of clusters or farms of servers acting as one large
resource pool.
By using these mechanisms, some of the traditional capacity management challenges in
computing environments can be addressed. In the one app, one physical server era for
commodity platforms, you normally had to provision for the peak demand of an application.
If the normal level of activity was considerably lower than that of a peak, a majority of the
resources sat unused most of the time. To make matters worse, the peaks could be weeks or
months apart depending on the cyclicality of events. In a virtualized environment, the extra
headroom needed to cover for peak usage can be shared among multiple different virtual
machines and across several physical hosts, all this in a way that is totally transparent to
the applications running inside the virtual machines. Combined with demand management
activities aimed at optimizing the timing of the peaks, dynamic resource scheduling can be
exceptionally effective at fighting excessive over-provisioning.
But are server virtualization and dynamic resources the silver bullet that will remove the need
for capacity management altogether? Not really, in the next section we will discuss why.
2 of 8 Capacity Management is Crucial in Virtualized Environments
White Paper
Copyright 2010 TeamQuest Corporation. All Rights Reserved.
