Extracting Resource Allocation Data for Chargeback in a HP Virtual Server Environment for HP Integrity Servers

workload
A
CPU Cores
workload
B
workload
C
Server X
Asset/ Server
Cost Recovery
Workload / Service
Provider
Server X
Cost to IT $Y
workload A, 16 % of Y
workload B, 33 % of Y
workload C, 50 % of Y
Users
Workload A, 1 Unit x $Rate
Workload B, 2 Units x $Rate
Workload C, 3 Units x $Rate
?
Requires server
inventory and
configuration
information as
well as workload
usage
Only requires
workload usage
information in
absolute units
such as CPU
cores
Figure 4 Server or Service focused reporting
For these reasons, this paper describes how to obtain the list of physical servers in the environment
and how to identify which workloads are associated with each.
Usage Parameters: Utilization, Allocation, and Guarantees
The second policy consideration that impacts the choice of parameters is the form of the resource
usage parameter. Of the three parameters described here, some are used more often than, but under
certain circumstances any of them might be chosen by IT.
• Utilization usage: This is the actual measured usage of resources by a user. Utilization is the
most intuitive and potentially most commonly used form of usage. Utilization is also the most
useful form for future optimization of resource allocation and planning.
• Allocated usage: For some server and virtualization technologies, resources can be allocated
to users in greater increments than they actually consume. As an implementation detail of IT
and “overhead,” avoid passing these allocations on to the end-user consumer. On the other
hand, IT might need to assign all costs of assets they deploy to the pool of users who use
them. In this case, allocated usage can represent a more direct metric of the costs of resource
deployment.
• Guaranteed usage: This is the amount of resources guaranteed to be available to the user
independent of how much the user actually consumes. When a guarantee exists, it can
represent the value that determines the sizing of the infrastructure and represents the critical
metric of what the service is costing IT to implement.
To illustrate these different forms of usage and why IT might choose one over another in different
situations, consider the following analogy to a member of a tennis club illustrated in Figure 5. If the
member is guaranteed a certain number of hours of court time per month, the club is likely to be very
sensitive to the number of members and associated guarantees relative to the number of courts and
the total schedulable court time. This factor is important for determining charges. If the courts are
allocated in hour-long blocks and the member does not use an entire hour of reserved time, the club
6