Machine Learning Resource Optimization Enabled by Cross Layer Monitoring

In this paper, we introduce a novel architecture and its open-source implementation that exploits the monitoring data from heterogeneous resources and uses them to train machine learning models, which can be used for dynamic resource management optimization. The existence of such a solution is extremely important for Service Providers (SP) as it can lead to the optimal use of their physical and virtual infrastructures avoiding potential waste of resources due to overdesign while at the same time it can ensure that the required Quality of Service (QoS) levels are met. The proposed solution is validated in two real-life services showing very good accuracy in predicting the required resources in both cases for a large number of operational scenarios.