MALIBOO: When Machine Learning meets Bayesian Optimization

Abstract

Bayesian optimization (BO) is an efficient method for finding optimal cloud computing configurations for several types of applications. On the other hand, Machine Learning (ML) methods can provide useful knowledge about the application at hand thanks to their predicting capabilities. In this paper, we propose a hybrid algorithm that is based on BO and integrates elements from ML techniques, to find the optimal configuration of time-constrained recurring jobs executed in cloud environments. The algorithm is tested by considering edge computing and Apache Spark big data applications. The results we achieve show that our approach reduces the amount of unfeasible executions up to 2-3 times with respect to state-of-the-art techniques.