Prediction-Based Power Oversubscription in Cloud Platforms

TL;DR

This paper proposes a prediction-based approach to increase power oversubscription in cloud datacenters, specifically in Microsoft Azure, by leveraging workload predictions to safely double oversubscription levels.

Contribution

It introduces a novel workload prediction method for criticality-aware power management, enabling higher oversubscription with minimal impact on performance.

Findings

Achieved 2x increase in oversubscription

Maintained performance of critical workloads

Demonstrated effectiveness in Microsoft Azure infrastructure

Abstract

Datacenter designers rely on conservative estimates of IT equipment power draw to provision resources. This leaves resources underutilized and requires more datacenters to be built. Prior work has used power capping to shave the rare power peaks and add more servers to the datacenter, thereby oversubscribing its resources and lowering capital costs. This works well when the workloads and their server placements are known. Unfortunately, these factors are unknown in public clouds, forcing providers to limit the oversubscription so that performance is never impacted. In this paper, we argue that providers can use predictions of workload performance criticality and virtual machine (VM) resource utilization to increase oversubscription. This poses many challenges, such as identifying the performance-critical workloads from black-box VMs, creating support for criticality-aware power management, and increasing oversubscription while limiting the impact of capping. We address these challenges for the hardware and software infrastructures of Microsoft Azure. The results show that we enable a 2x increase in oversubscription with minimum impact to critical workloads.