SOM-based behavioral analysis for virtualized network functions

Abstract

In this paper, we propose a mechanism based on Self-Organizing Maps for analyzing the resource consumption behaviors and detecting possible anomalies in data centers for Network Function Virtualization (NFV). Our approach is based on a joint analysis of two historical data sets available through two separate monitoring systems: system-level metrics for the physical and virtual machines obtained from the monitoring infrastructure, and application-level metrics available from the individual virtualized network functions. Experimental results, obtained by processing real data from one of the NFV data centers of the Vodafone network operator, highlight some of the capabilities of our system to identify interesting points in space and time of the evolution of the monitored infrastructure.

XPySom: High-Performance Self-Organizing Maps

“…Note that we have also run XPySom on an industrial dataset -provided by Vodafone -regarding data-center metrics for network function virtualization, that has been mentioned in our previous works [13], [14]. From a preliminary experimentation, the conclusions in terms of performance gain of XPySom compared to SomoClu are fundamentally the same as shown in this paper.…”

“…Indeed, they are designed for mapping highdimensional data into a lower-dimensional space (e.g., 2D) that is better interpretable by human perception and easier to treat computation-wise, while preserving the topology and distribution of the original data at cluster-level. Given their ability to yield a data distribution in the target domain that faithfully reflects the observed relationships in the original space, SOMs have achieved remarkable results in many application fields like: image processing [4], [5], industrial data processing [6], [7], data visualization [8]- [10], pattern recognition [11], [12], anomaly detection in NFV infrastructures [13], [14].…”

Forecasting Operation Metrics for Virtualized Network Functions

“…For this reason, Machine Learning (ML) techniques have been gaining momentum as key technologies accompanying enterprises operating in pretty much any business domain. In networking, these techniques have been successfully applied in NFV infrastructures for anomaly detection [3], [4], behavioral pattern analysis [5], [6] as well as resource demand estimations [7], [8]. In particular, Deep Learning (DL) methods are among the techniques that are receiving increasing attention from both research and industry.…”