Optimize Server Utilization for Energy Efficiency and Embodied Carbon

Energy Proportionality of Servers

Servers are most efficient at high utilization rates because of Energy Proportionality of Servers and Dynamic Range.

Embodied Carbon of Servers

Every server has an embodied carbon footprint. If a server is under-utilized, it has a bigger footprint per work unit than a high-utilized server.

Overall Energy Reduction Potentials

Comparison between change in IT utilization and change in PUE

Intel Gold 6230 Xeon processor (The Green Grid database):

40% utilization: 286 watts (W)
20% utilization: 267 W
→ Reduction: 45%

Source: Dietrich, J., & Lawrence, A. (2022). Reducing the energy footprint (Digital Infrastructure Sustainability – a Manager’s Guide, p. 8). Uptime Institute.

→ Better server utilization is often much more important than increasing the PUE.

Goal

What capacity utilization should be aimed for in order to achieve the highest possible energy efficiency and to be able to handle peak loads well?

Dropbox

At Dropbox, we aim to reach 85% capacity, a sweet spot for usage that’s efficient yet still gives us extra capacity to handle power spikes.

Strategies

What strategies exist to achieve a high server utilization?

Shutdown Zombie Servers

25% of all physical servers and 30% of all virtual servers are comatose (Koomey & Anthesis, 2017).

→ Zombie servers reserve resources that are not used, that lead to underutilization.

Right-sizing

amazonwebservices.2023.sustainabilitypillaraws (pg. 9)

Maximize utilization: Right-size workloads and implement efficient design to ensure high utilization and maximize the energy efficiency of the underlying hardware. Two hosts running at 30% utilization are less efficient than one host running at 60% due to baseline power consumption per host. At the same time, eliminate or minimize idle resources, processing, and storage to reduce the total energy required to power your workload.

Auto-scaling

Kubernetes Auto-Scaling