FIRM: An Intelligent Fine-Grained Resource Management Framework for SLO-Oriented Microservices

TL;DR

FIRM is an intelligent resource management framework that uses machine learning and telemetry data to dynamically allocate resources in microservices, significantly reducing SLO violations and tail latencies.

Contribution

FIRM introduces a novel adaptive framework that detects, localizes, and mitigates resource contention in microservices using online data and machine learning techniques.

Findings

Reduces SLO violations by up to 16x

Decreases overall CPU requests by up to 62%

Reduces tail latencies by up to 11x

Abstract

Modern user-facing latency-sensitive web services include numerous distributed, intercommunicating microservices that promise to simplify software development and operation. However, multiplexing of compute resources across microservices is still challenging in production because contention for shared resources can cause latency spikes that violate the service-level objectives (SLOs) of user requests. This paper presents FIRM, an intelligent fine-grained resource management framework for predictable sharing of resources across microservices to drive up overall utilization. FIRM leverages online telemetry data and machine-learning methods to adaptively (a) detect/localize microservices that cause SLO violations, (b) identify low-level resources in contention, and (c) take actions to mitigate SLO violations via dynamic reprovisioning. Experiments across four microservice benchmarks demonstrate that FIRM reduces SLO violations by up to 16x while reducing the overall requested CPU limit by up to 62%. Moreover, FIRM improves performance predictability by reducing tail latencies by up to 11x.