Quadrant 2 - IT Assets

The second group of KPI’s is that of IT hardware asset efficiency. Though the variables measured differ between the various architectures, all four are defined as usage percentage of the current capabilities divided by a reference capability. This definition again makes for KPI’s that are flexible and can be tuned to the particular environment under study.

The goal is to determine the balance between installed capacity and used capacity. Most IT hardware uses energy regardless the amount of work.

The formula to calculate the asset efficiency is

asset efficiency = (current capacity X average
utilization) / reference capacity

A basic example in a number of steps:

  1. The disks that are in use in the storage unit are 1TB 7,200 rpm SATA disks.
  2. The average capacity utilization of the storage unit is 35%.
  3. The current/modern size of a 7.200 rpm SATA disk is 2TB.
  4. The asset efficiency is (1TB * 30%) / 2TB = 15%.

This means that the first step in determining the asset efficiency of the storage unit results in 15%. Because if you would buy the storage unit today, you could buy it with 2TB disks instead of 1TB disks. And because you are using 35% of its current capacity, why didn’t you buy less capacity initially and add more, with larger disks, when needed? The next step would be to add the power draw of the selected components to the math. A larger or newer disk could use less energy than the ones that are installed. And if it uses the same about of power, it is less power per capacity unit (TB in this example).

To be able to achieve the highest score you need 100% utilized capacity on the most modern hardware. Having 100% utilized capacity is not desired, so adjust the average utilization parameter to reflect the utilization of your desired maximum. If 80% is the maximum, then adjust the 35% from the example by multiplying it to 100/80 which results in 43,75. So, the score is 43,75 instead of 35 when taking into account that 80% utilized capacity means “full”. Set yours at a challenging level per KPI or even for multiple types of hardware in each KPI to prevent easy scoring but to achieve realistic scoring.

These KPI’s are defined:

Processor architectures like mainframes, supercomputers and others are not in the model by default. If a data center has a significant size of other architectures than x86 and RISC, then add or replace those specific architectures in the model. You can adapt it to your own need.