Optimizing the cloud? Don't train models. Build oracles!

TL;DR

This paper introduces cloud oracles as an accurate, explainable alternative to machine learning for optimizing cloud configurations, especially for parametric convex problems, with promising experimental results and future research directions.

Contribution

It presents a novel approach using cloud oracles for cloud configuration optimization, offering guaranteed accuracy and explainability over traditional machine learning methods.

Findings

Proposes cloud oracles as an effective optimization tool.

Demonstrates efficacy through experimental evidence.

Highlights potential for expanding applicability in future research.

Abstract

We propose cloud oracles, an alternative to machine learning for online optimization of cloud configurations. Our cloud oracle approach guarantees complete accuracy and explainability of decisions for problems that can be formulated as parametric convex optimizations. We give experimental evidence of this technique's efficacy and share a vision of research directions for expanding its applicability.