Workload Similarity Analysis using Machine Learning Techniques

TL;DR

This paper introduces a machine learning-based method to measure workload similarity using performance telemetry data, aiding in workload characterization and proxy creation, with demonstrated accuracy on benchmark workloads.

Contribution

It presents a novel approach leveraging machine learning to analyze workload telemetry data for similarity measurement, which was not previously explored in this context.

Findings

High accuracy in workload similarity measurement

Effective characterization of benchmark workloads

Potential for creating proxy workloads

Abstract

Finding the similarity between two workload behaviors is helpful in 1. creating proxy workloads 2. characterizing an unknown workload's behavior by matching its behavior against known workloads. In this article, we propose a method to measure the similarity between two workloads using machine learning-based analysis of the performance telemetry data collected for the execution runs of the two workloads. We also demonstrate the accuracy of the technique by measuring the similarity between a variety of know benchmark workloads.