Overview
ISO/IEC 30134-5:2017 defines IT Equipment Utilization for servers (ITEUsv) as a Key Performance Indicator (KPI) to quantify average server utilization in a data centre. The standard specifies a conceptual definition, measurement approach, reporting rules and practical guidance to use CPU utilization as a proxy for server activity. ITEUsv is intended to help data centre owners/operators improve the aggregate energy efficiency of servers through consolidation, virtualization and operational changes.
Key topics and technical requirements
- Definition: ITEUsv is the average CPU utilization ratio of all servers (or a designated group) in a data centre.
- Per‑time formula: At time t, ITEUsv(t) = (Σ CUSi(t)) / N, where CUSi(t) is CPU utilization for server i and N is the number of running servers.
- Annual measurement: Monitor ITEUsv(t) continuously for one year and compute the annual average ITEUsv. Measurement intervals (e) must be between 1 minute and 1 hour; choose interval based on workload dynamics.
- Multicore/CPU handling: For multicore systems, CUSi(t) is the average utilization across cores; for multiple CPUs, the average across CPUs.
- Exclusions:
- Servers where most computation is done by auxiliary processing accelerators (APAs) are out of scope.
- Servers without applicable CPU monitoring tools should be excluded from the measurement group.
- Servers in “power OFF” state are subtracted when calculating ITEUsv.
- Reporting: Record annual average, peak ITEUsv(t) and peak duration metrics to support target-setting and operational planning.
- Limitations: ITEUsv relies on CPU utilization as a proxy - consider memory, I/O, workload changes and differing OS/CPU measurement semantics when interpreting results.
Practical applications and users
- Who uses it: data centre operators, hosting/cloud providers, sustainability and capacity planners, IT operations and performance engineers.
- Use cases:
- Track and improve server energy effectiveness by increasing average utilization (via virtualization, consolidation, workload scheduling).
- Validate the impact of operational changes (server provisioning, autoscaling policies, power management).
- Set internal targets (annual average and acceptable peak durations) and monitor trends year-over-year.
- Operational guidance:
- Choose an appropriate sampling interval (1 min–1 h) reflecting workload variability.
- Monitor continuously for a year for representative results.
- Use peak and duration data to avoid unacceptable performance impact when raising average utilization.
Related standards
- ISO/IEC 30134-1 - Overview and general requirements for data centre KPIs.
- ISO/IEC 30134-2 - PUE (Power Usage Effectiveness) for infrastructure energy efficiency.
- ISO/IEC 30134-4 - Addresses capability/capacity aspects complementary to utilization.
Keywords: ISO/IEC 30134-5:2017, ITEUsv, data centre KPI, server utilization, CPU utilization, energy efficiency, PUE, server consolidation.