Predicting Response Time-Related Quality-of-Service Outages of PaaS Cloud Applications by Machine Learning

Creators: Schedel, Angela and Brune, Philipp
Title: Predicting Response Time-Related Quality-of-Service Outages of PaaS Cloud Applications by Machine Learning
Item Type: Book Section
Page Range: pp. 155-165
Date: 27 September 2017
Divisions: Informationsmanagement
Abstract: For customers running their applications on Platform-as-a-Service (PaaS) cloud environments it is important to ensure the Quality-of-Service (QoS) of their applications. Knowing in advance if and when a potential problem is likely to occur allows the application owner to take appropriate countermeasures. Therefore, predictive analytics using machine learning could allow to be alerted in advance about potential upcoming QoS outages. In this context, mainly Infrastructure-as-a-Service (IaaS) or Software-as-a-Service (SaaS) have been studied in the literature so far. Studies about predicting QoS outages for the Platform-as-a-Service (PaaS) service model are sparse. Therefore, in this paper an approach for predicting response-time-related QoS outages of web services running in a PaaS cloud environment is presented. The proposed solution uses the open source Apache Spark platform in combination with MLib and binary classification by the naive Bayes algorithm. The approach is evaluated by using test data from a social app backend web service. The results indicate that it is feasible in practice.
Citation:

Schedel, Angela and Brune, Philipp (2017) Predicting Response Time-Related Quality-of-Service Outages of PaaS Cloud Applications by Machine Learning. In: Mobile, Secure, and Programmable Networking: Third International Conference, MSPN 2017, Paris, France, 2017. Lecture Notes in Computer Science (LNCS), 10566 . Springer, Cham, pp. 155-165. ISBN 978-3-319-67807-8

Actions (login required)

View Item View Item