A Machine Learning-based Framework for Building Application Failure Prediction Models
Alessandro Pellegrini, Pierangelo Di Sanzo, and Dimiter R. Avresky
Published in: Proceedings of the 20th IEEE Workshop on Dependable Parallel, Distributed and Network-Centric Systems
Download PDF
Abstract:
In this paper, we present the Framework for building Failure Prediction Models (F2PM), a Machine Learning-based Framework to build models for predicting the Remaining Time to Failure (RTTF) of applications in the presence of software anomalies. F2PMuses measurements of a number of system features in order to create a knowledge base, which is then used to build prediction models. F2PM is application-independent, i.e. It solely exploits measurements of system-level features. Thus, it can be used in differentiated contexts, without the need for any manual modification or intervention to the running applications. To generate optimized models, F2PM can perform a feature selection to identify, among all the measured system features, which have a major impact in the prediction of the RTTF. This allows to produce different models, which use different set of input features. Generated models can be compared by the user by using a set of metrics produced by F2PM, which are related to the model prediction accuracy, as well as to the model building time. We also present experimental results of a successful application of F2PM, using the standard TPC-W e-commerce benchmark.
BibTeX Entry:
author = {Pellegrini, Alessandro and Di Sanzo, Pierangelo and Avresky, Dimiter R.},
booktitle = {Proceedings of the 20th IEEE Workshop on Dependable Parallel, Distributed and Network-Centric Systems},
title = {A Machine Learning-based Framework for Building Application Failure Prediction Models},
year = {2015},
month = may,
pages = {1072--1081},
publisher = {IEEE Computer Society},
series = {DPDNS},
doi = {10.1109/IPDPSW.2015.110},
location = {Hyderabad, India}
}