An Interference-Aware Application Classifier Based on Machine Learning to Improve Scheduling in Clouds

ML-driven classification scheme for dynamic interference-aware resource scheduling in cloud infrastructures

“…After trained, the proposed classifier receives monitored metrics from applications and dynamically defines their interference levels thresholds for each resource. This application classifier was first introduced in [49] in its static variant, and here we improve this work by introducing a dynamic version, where we apply the classifier several times during the execution of an application to better react to workload variations and possible changes in its interference levels.…”

Using Machine Learning Techniques to Classify the Interference of HPC applications in Virtual Machines with Uncertain Data

“…Meyer et al [Meyer et al 2020] proposed a two-phase interference-aware classifier. The first phase uses a classification technique to determine in which of five possible classes (memory, CPU, disk, network, or cache) the object can suffer from interference.…”

“…To tackle this problem, some works such as [Alves and de Assumpção Drummond 2017, Ludwig et al 2019, Ren et al 2019, Zacarias et al 2019, Meyer et al 2020 proposed models to predict the level of interference suffered by a set of applications allocated to the same environment. Although those proposed models presented satisfactory results, some rely on normalized application access rates to shared resources, which are hard information to obtain, especially for end-users.…”