The Performance Indicator (PI) is a set of metrics developed by The Green Grid for assessing data center power and cooling effectiveness.
The PI is based on and expands the power usage effectiveness (PUE) metric that has been in use since 2007. Organizations can use PI to evaluate their data centers' cooling performance, and to predict future thermal states. PI integrates variables such as data center capacity, sustainability and availability.
The three metrics of the PI
The PI measurements are based on three metrics:
- PUE ratio -- Assesses the energy efficiency of a data center based on the PUE metric
- IT Thermal Conformance – Assesses how efficiently equipment is cooled during normal IT operations
- IT Thermal Resilience – Assesses how efficiently equipment is cooled during maintenance periods or failures
PI measures each of these metrics individually, and determines how a change in one metric affects the others.
The four levels of the PI
There are four levels of PI measurement, ranging from the most basic to the most precise. The amount of data an organization can collect about their IT infrastructure dictates the level of PI measurement it can use. Those who can collect more data, through temperature monitoring sensors, data center infrastructure management (DCIM) tools and other means, can achieve a higher-level and more accurate PI level than those who cannot. Facilities can choose which level works best for them based on their infrastructure and needs.Content Continues Below
The PI levels include:
- Level 1 – The most basic PI assessment, with temperature monitoring sensors placed at the server rack level. The Level 1 measurement can only assess a facility's current state.
- Level 2 – Builds on the Level 1 measurement to achieve a more accurate assessment, but still only reflects a facility's current state. The Green Grid recommends organizations use a DCIM system to make Level 2 more attainable.
- Level 3 -- The Level 3 PI measures adds basic predictive modeling capabilities to help IT predict a facility's future cooling efficiency, in addition to its current state.
- Level 4 – The most sophisticated PI assessment, allowing for the most accurate prediction of future performance, based on 3D simulations, in addition to current states.