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

摘要: For customers running their applications on Platform-as-a-Service (PaaS) cloud environments it is important to ensure the Quality-of-Service (QoS) of applications. Knowing in advance if and when a potential problem likely occur allows application owner take appropriate countermeasures. Therefore, predictive analytics using machine learning could allow be alerted about upcoming QoS outages. In this context, mainly Infrastructure-as-a-Service (IaaS) or Software-as-a-Service (SaaS) have been studied literature so far. Studies predicting outages for service model are sparse. paper an approach response-time-related web services PaaS environment presented. The proposed solution uses open source Apache Spark platform combination with MLib binary classification by naive Bayes algorithm. evaluated test data from social app backend service. results indicate that feasible practice.